This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
perl5.git
9 years agolib/h2ph.t to test generated t/_h2ph_pre.ph instead of the system one
Petr Písař [Tue, 17 Feb 2015 12:11:00 +0000 (13:11 +0100)]
lib/h2ph.t to test generated t/_h2ph_pre.ph instead of the system one

The lib/h2ph.t test executes a t/lib/h2ph.pht which requires
'_h2ph_pre.ph'.  This should find and exercise generated t/_h2ph_pre.ph
file. However, it found a loaded _h2ph_pre.ph from system because the
interpreter has the './' directory after the system paths in the @INC by
default.

This patch adds '-I./' to the runperl() invocation to prefer the
_h2ph_pre.ph generated at build time.

Signed-off-by: Petr Písař <ppisar@redhat.com>
9 years agofix some spurious PERL_UNUSED_ARG/VAR() usage
Jasmine Ngan [Tue, 24 Feb 2015 14:37:17 +0000 (14:37 +0000)]
fix some spurious PERL_UNUSED_ARG/VAR() usage

The 'output_warning' param is actually used, and 'items' is not a function
argument but a local variable.

9 years agosilence some compiler warnings.
David Mitchell [Tue, 24 Feb 2015 12:12:21 +0000 (12:12 +0000)]
silence some compiler warnings.

(Some compilers have strange ideas about the signedness of the bool type
verses the signedness of a boolean expression like a == b).

9 years agoUpdate Getopt-Long to CPAN version 2.45
Chris 'BinGOs' Williams [Mon, 23 Feb 2015 19:41:50 +0000 (19:41 +0000)]
Update Getopt-Long to CPAN version 2.45

  [DELTA]

Changes in version 2.45
-----------------------

* Revert behaviour of the combination of <> and pass_through to what
  is was in 2.42.

  Thanks to Wilson Snyder to provide enhancements to the documentation
  to clarify why <> and pass_through generally is a bad idea.

  It is, however, very unfortunate to hear from users that their
  programs *depend* on the undocumented pre-2.43 behaviour.

9 years ago[perl #123801] Stop s##[}#e from crashing
Father Chrysostomos [Sun, 22 Feb 2015 22:34:42 +0000 (14:34 -0800)]
[perl #123801] Stop s##[}#e from crashing

The lexer normally turns s##...#e into

PMFUNC '(' WORD '/' DO '{' ... ';' '}' ')'

If you have [} inside the replacement, that becomes '[' ';' '}'.  When
the parser gets to the second semicolon, it pops scopes to try to
recover from the syntax error, and in so doing it exits the inner lex-
ing scope that was set up for the substitution.

When that happens, the second '}' is already on the pending token
stack.  Since we set the lexing state to LEX_KNOWNEXT when there is a
pending token (though we may not have to; see 7aa8cb0dec1), we have to
record a pending state as well, so we know what to set the state back
to.  That pending state is not localised, and, in this case, was set
before the scopes were popped.

So we end up in the outermost lexing scope, with the lexing state set
to LEX_INTERPEND.

Inside an inner lexing scope, PL_linestr is of type PVIV, with the IVX
field used to hold extra information about the type of quote.  In the
main lexing scope, PL_linestr is an SVt_PV with no IVX field.

If the lexing state is LEX_INTERPanything, it is assumed that
PL_linestr has an IVX field, which is not the case here, so we fail an
assertion or crash.

The safest pre-5.22 solution is to check the type of PL_linestr before
reading IVX.

9 years agoparser.h: Make nexttoke unsigned
Father Chrysostomos [Sun, 22 Feb 2015 19:26:15 +0000 (11:26 -0800)]
parser.h: Make nexttoke unsigned

This is to avoid a compiler warning at toke.c:1912 (the assertion in
S_force_next).  The only values this member can contain are 0 to 5, so
this change is safe.  (I’ll probably change it to U8 after 5.22.)

9 years agoTest-Simple PR#494 was merged; t/Legacy/exit.t is no longer customized in blead
Steve Hay [Sun, 22 Feb 2015 15:45:28 +0000 (15:45 +0000)]
Test-Simple PR#494 was merged; t/Legacy/exit.t is no longer customized in blead

9 years agoRemove three Test-Simple files which are no longer in Test-Simple on CPAN
Steve Hay [Sun, 22 Feb 2015 15:38:18 +0000 (15:38 +0000)]
Remove three Test-Simple files which are no longer in Test-Simple on CPAN

9 years agoblead doesn't include cpan/perlfaq/inc/perlfaq.tt
Steve Hay [Sun, 22 Feb 2015 15:32:58 +0000 (15:32 +0000)]
blead doesn't include cpan/perlfaq/inc/perlfaq.tt

9 years agoPod::Parser is no longer customized in blead, since commit 534577b24e
Steve Hay [Sun, 22 Feb 2015 15:29:06 +0000 (15:29 +0000)]
Pod::Parser is no longer customized in blead, since commit 534577b24e

9 years ago[perl #123893] Fix hang with "@{"
Father Chrysostomos [Sun, 22 Feb 2015 06:24:20 +0000 (22:24 -0800)]
[perl #123893] Fix hang with "@{"

Commit v5.21.8-320-ge47d32d stopped code interpolated into quote-like
operators from reading more lines of input, by making lex_next_chunk
ignore the open filehandle and return false.  That causes this block
under case 0 in yylex to loop:

    if (!lex_next_chunk(fake_eof)) {
CopLINE_dec(PL_curcop);
s = PL_bufptr;
TOKEN(';'); /* not infinite loop because rsfp is NULL now */
    }

(rsfp is not null there.)  This commit makes it check for quote-like
operators above, in the same place where it checks whether the file is
open, to avoid falling through to this code that can loop.

This changes the syntax errors for a couple of cases recently added
to t/op/lex.t, though I think the error output is now more consis-
tent overall.

9 years agoperlrebackslash: Amplify and correct \b{sb}, \b{wb}
Karl Williamson [Sat, 21 Feb 2015 19:16:40 +0000 (12:16 -0700)]
perlrebackslash: Amplify and correct \b{sb}, \b{wb}

9 years agoperlrecharclass: \R could be inside a [...]
Karl Williamson [Sat, 21 Feb 2015 19:12:35 +0000 (12:12 -0700)]
perlrecharclass: \R could be inside a [...]

But we haven't made it so.  So remove the text saying that it can't.
For a couple of releases now, we have allowed something that can match
more than one character to be inside a non-complemented bracketed
character class, so this could too, should we choose to someday.

9 years ago\s matching VT is no longer experimental
Karl Williamson [Fri, 20 Feb 2015 18:10:05 +0000 (11:10 -0700)]
\s matching VT is no longer experimental

This was experimentally introduced in 5.18, and no issues were raised,
except that it got us to thinking and spurred us to stop allowing $^X,
where 'X' is a non-printable control character, and that change caused
some issues.

9 years agocharclass_invlists.h: Add regen entry
Karl Williamson [Fri, 20 Feb 2015 18:09:41 +0000 (11:09 -0700)]
charclass_invlists.h: Add regen entry

This missing entry is one used by t/porting/regen.t to see if the
contents are up-to-date.  I don't know why it didn't get added earlier,
and why there aren't failures except apparently on my machine due to
it's not being there.  I thought I took great care in getting it right.

9 years agoperlunitut, perlreref: Nits
Karl Williamson [Fri, 20 Feb 2015 17:12:52 +0000 (10:12 -0700)]
perlunitut, perlreref: Nits

Spotted by Tom Christiansen

9 years agoBump $VERSION in CPAN modules changed since 5.21.9
Steve Hay [Sat, 21 Feb 2015 18:17:38 +0000 (18:17 +0000)]
Bump $VERSION in CPAN modules changed since 5.21.9

9 years agoUpgrade CPAN from version 2.05 to 2.09-TRIAL
Steve Hay [Sat, 21 Feb 2015 18:06:25 +0000 (18:06 +0000)]
Upgrade CPAN from version 2.05 to 2.09-TRIAL

9 years agoPorting/bump-perl-version: No need to try other maps after one has worked
Steve Hay [Thu, 19 Feb 2015 08:42:46 +0000 (08:42 +0000)]
Porting/bump-perl-version: No need to try other maps after one has worked

9 years agoModule-CoreList-5.20150220 is now on CPAN
Steve Hay [Sat, 21 Feb 2015 16:44:31 +0000 (16:44 +0000)]
Module-CoreList-5.20150220 is now on CPAN

9 years agoOne $VERSION bump missed by 2f553ae1e1
Steve Hay [Sat, 21 Feb 2015 16:11:33 +0000 (16:11 +0000)]
One $VERSION bump missed by 2f553ae1e1

9 years agoUpdate Getopt-Long to CPAN version 2.44
Chris 'BinGOs' Williams [Sat, 21 Feb 2015 13:08:08 +0000 (13:08 +0000)]
Update Getopt-Long to CPAN version 2.44

  [DELTA]

Changes in version 2.44
-----------------------

* Be gentle to the user community and turn fix 92462 into a warning.
  This may be changed back to error in some future release.

9 years agoadd missing infnan context under thread/quadmath builds
Lukas Mai [Sat, 21 Feb 2015 10:29:45 +0000 (11:29 +0100)]
add missing infnan context under thread/quadmath builds

9 years agotweak version number of Module::CoreList
Sawyer X [Sat, 21 Feb 2015 07:27:54 +0000 (08:27 +0100)]
tweak version number of Module::CoreList

9 years agoadd some dummy entries to Module::CoreList for 5.21.10
Sawyer X [Sat, 21 Feb 2015 06:54:55 +0000 (07:54 +0100)]
add some dummy entries to Module::CoreList for 5.21.10

9 years agoadded link to announcement
Sawyer X [Sat, 21 Feb 2015 06:44:28 +0000 (07:44 +0100)]
added link to announcement

9 years agonew delta for 5.21.10
Sawyer X [Sat, 21 Feb 2015 06:15:31 +0000 (07:15 +0100)]
new delta for 5.21.10

9 years agoadd my epigraph, tick 5.21.9 release
Sawyer X [Sat, 21 Feb 2015 06:06:54 +0000 (07:06 +0100)]
add my epigraph, tick 5.21.9 release

9 years agoadd new release to perlhist v5.21.9
Sawyer X [Sat, 21 Feb 2015 04:03:37 +0000 (05:03 +0100)]
add new release to perlhist

9 years agofinalize perldelta and Module::CoreList updates
Sawyer X [Sat, 21 Feb 2015 03:39:03 +0000 (04:39 +0100)]
finalize perldelta and Module::CoreList updates

9 years agofix typos
Sawyer X [Sat, 21 Feb 2015 03:34:46 +0000 (04:34 +0100)]
fix typos

9 years agoUpdate Module::CoreList for 5.21.9
Sawyer X [Sat, 21 Feb 2015 02:03:13 +0000 (03:03 +0100)]
Update Module::CoreList for 5.21.9

9 years agopodcheck: accept links to autobox
Ricardo Signes [Sat, 21 Feb 2015 01:14:28 +0000 (20:14 -0500)]
podcheck: accept links to autobox

9 years agoUpdate perldelta
Sawyer X [Sat, 21 Feb 2015 00:10:25 +0000 (01:10 +0100)]
Update perldelta

9 years agopp_pack.c: Silence compiler warning
Karl Williamson [Fri, 20 Feb 2015 16:27:16 +0000 (09:27 -0700)]
pp_pack.c: Silence compiler warning

This was introduced by 9df874cdaa2f196cc11fbd7b82a85690c243eb9f
in changing the name of some static functions.  I didn't realize at the
time that the function was defined in embed.fnc, as none of the others
are, and it was always called with the S_ prefix form.  Nor did I notice
the compiler warnings.

It turns out that the base name of this function is the same as a public
function, so I've renamed it to have prefix 'S_my_'.

9 years agoupdate perldelta for some modules and pragmas upgrades
Sawyer X [Fri, 20 Feb 2015 07:28:20 +0000 (08:28 +0100)]
update perldelta for some modules and pragmas upgrades

9 years agoperldelta for e47d32dcd and d27f4b91 / #123712
Father Chrysostomos [Fri, 20 Feb 2015 06:40:29 +0000 (22:40 -0800)]
perldelta for e47d32dcd and d27f4b91 / #123712

9 years agoperldelta for parser crashes
Father Chrysostomos [Fri, 20 Feb 2015 06:36:53 +0000 (22:36 -0800)]
perldelta for parser crashes

eabab8bcc
488bc5795
fb0c7c3c4
3c47da3c2
eea89386b
1f7c3e7c6
7aa8cb0de

9 years agoperldelta for df442ae4d / #123466
Father Chrysostomos [Fri, 20 Feb 2015 06:31:24 +0000 (22:31 -0800)]
perldelta for df442ae4d / #123466

9 years agoperldelta for 2f465e08e / #123652
Father Chrysostomos [Fri, 20 Feb 2015 06:26:25 +0000 (22:26 -0800)]
perldelta for 2f465e08e / #123652

9 years agoregcomp.c: Add assertion
Karl Williamson [Fri, 20 Feb 2015 04:03:44 +0000 (21:03 -0700)]
regcomp.c: Add assertion

9 years agoAdd \b{sb}
Karl Williamson [Fri, 20 Feb 2015 02:42:33 +0000 (19:42 -0700)]
Add \b{sb}

9 years agoregen/mk_invlists.pl: Add tables for \b{sb}
Karl Williamson [Fri, 20 Feb 2015 02:37:47 +0000 (19:37 -0700)]
regen/mk_invlists.pl: Add tables for \b{sb}

This single line addition generates a very confused diff listing for the
generated file.

9 years agoAdd qr/\b{wb}/
Karl Williamson [Thu, 19 Feb 2015 22:19:08 +0000 (15:19 -0700)]
Add qr/\b{wb}/

9 years agoregen/mk_invlists.pl: Add tables for \b{wb}
Karl Williamson [Thu, 19 Feb 2015 21:54:54 +0000 (14:54 -0700)]
regen/mk_invlists.pl: Add tables for \b{wb}

This single line change generates very a confused diff listing for the
generated file, so is kept separate form the other \b{wb} commits.

9 years agoRemove obsolete macros/tables for \X
Karl Williamson [Thu, 19 Feb 2015 20:21:22 +0000 (13:21 -0700)]
Remove obsolete macros/tables for \X

A previous commit changed how \X is implemented, and now we don't need
these anymore.

9 years agoregexec.c: White-space only
Karl Williamson [Mon, 9 Feb 2015 06:04:18 +0000 (23:04 -0700)]
regexec.c: White-space only

Add indentation, as the previous commit added a switch{} around these.

9 years agoAdd qr/\b{gcb}/
Karl Williamson [Tue, 17 Feb 2015 22:03:32 +0000 (15:03 -0700)]
Add qr/\b{gcb}/

A function implements seeing if the space between any two characters is
a grapheme cluster break.  Afer I wrote this, I realized that an array
lookup might be a better implementation, but the deadline for v5.22 was
too close to change it.  I did see that my gcc optimized it down to
an array lookup.

This makes the implementation of \X go from being complicated to
trivial.

9 years agoregen/mk_invlists.pl: Add GCB tables
Karl Williamson [Wed, 18 Feb 2015 22:40:26 +0000 (15:40 -0700)]
regen/mk_invlists.pl: Add GCB tables

This will enable the next commit to add \b{gcb}.

I separated this out from that commit because the diff output here is is
very confused, not accurately showing the underlying changes.  Actually
two data structures are being added for every character set, and nothing
else changed.

9 years agoregen/mk_invlists.pl: Revamp #if generation
Karl Williamson [Thu, 19 Feb 2015 04:49:16 +0000 (21:49 -0700)]
regen/mk_invlists.pl: Revamp #if generation

This changes where the symbols are defined to a single file each.  This
may save text space, depending on the compiler.  The next commit will
cause this hdr to be included in more places, so it becomes more
important to do this.

At the same time this removes the guard for #ifndef PERL_IN_XSUB_RE.
The code now is executed regardless of that.  This is simpler, and
previously there might have been the possibility of uninitialized memory
being read, should re_comp.o be executed before recomp.o.

9 years agoregen/mk_invlists.pl: Add capability for line break props
Karl Williamson [Wed, 18 Feb 2015 22:29:09 +0000 (15:29 -0700)]
regen/mk_invlists.pl: Add capability for line break props

This is a partial implementation of a full inversion map generation
capability, which is why some code is indented more than necessary --
in the future there will be things that use that.  But this is
sufficient for 5.22.

This allows for the generation of tables to handle the Unicode line
breaking properties, like GCB and WB.  Future commits will actually use
this capability.

9 years agomktables: Refactor \X test so can be used for others
Karl Williamson [Fri, 20 Feb 2015 01:49:56 +0000 (18:49 -0700)]
mktables: Refactor \X test so can be used for others

The test file generated by mktables will soon contain other types of
breaks than \X.  This prepares for that by making a more general
function that can be the bottom level for each

9 years agoRemove another dead function in vms/vms.c.
Craig A. Berry [Fri, 20 Feb 2015 03:11:50 +0000 (21:11 -0600)]
Remove another dead function in vms/vms.c.

No use for the utf8_to_vtf7 function has appeared in the first
nine years of its existence so it's not earning its keep.

9 years agoFix some style issues in vms/vms.c.
Craig A. Berry [Fri, 20 Feb 2015 02:51:15 +0000 (20:51 -0600)]
Fix some style issues in vms/vms.c.

This normalizes function definitions and addresses the most
egregious of the Totally Forgot to Indent At All problems that are
rampant throughout this file.

For now I've punted on normalizing indents and tried to be
consistent with whatever local variant seems to be in use. There
are 2-, 3- (!), and 4-space indents, along with a liberal
sprinkling of tab characters, most (but not all) of which seem to
be intended to be equivalent to 8 spaces.

I considered normalizing indents with some like

   astyle -c --style=linux vms.c

but that produces 8400+ changed lines, which doesn't seem worth it.

9 years agoUpdate copyright in vms/vms.c.
Craig A. Berry [Thu, 19 Feb 2015 23:39:19 +0000 (17:39 -0600)]
Update copyright in vms/vms.c.

9 years agoregen podcheck.t db
Karl Williamson [Thu, 19 Feb 2015 20:25:20 +0000 (13:25 -0700)]
regen podcheck.t db

The RMG now has one more verbatim line longer than 79 columns

9 years agoconfigure.com does handle usedevel.
Craig A. Berry [Thu, 19 Feb 2015 01:10:19 +0000 (19:10 -0600)]
configure.com does handle usedevel.

So don't hard-code it to undef as was done in 1be1b388c18674b89.

9 years agoFix t/op/dump.t warning in skip case.
Craig A. Berry [Thu, 19 Feb 2015 01:06:29 +0000 (19:06 -0600)]
Fix t/op/dump.t warning in skip case.

When the whole test is skipped, the END block that does a chdir
to $start throws a warning because $start is not defined.  So
check for definedness.

9 years agoRandy Stauneris now a perl author
Tony Cook [Wed, 18 Feb 2015 23:08:58 +0000 (10:08 +1100)]
Randy Stauneris now a perl author

9 years agoFix minor code typos in perlunicook
Randy Stauner [Tue, 17 Feb 2015 14:43:18 +0000 (07:43 -0700)]
Fix minor code typos in perlunicook

9 years agoRMG: avoid saying "final" release; gets confused with BLEAD-FINAL
Ricardo Signes [Wed, 18 Feb 2015 21:11:08 +0000 (16:11 -0500)]
RMG: avoid saying "final" release; gets confused with BLEAD-FINAL

9 years agoutf8.c: Slight refactor of UTF-16 code
Karl Williamson [Tue, 17 Feb 2015 22:25:21 +0000 (15:25 -0700)]
utf8.c: Slight refactor of UTF-16 code

This eliminates a branch in the usual case, at the expense of an extra
one in the rarer case, which allows us to collapse some error condition
code.   It sprinkles some UNLIKELYs.

9 years agoperlre: Nits
Karl Williamson [Tue, 17 Feb 2015 21:48:32 +0000 (14:48 -0700)]
perlre: Nits

9 years agoperldelta: Fix typo
Karl Williamson [Wed, 18 Feb 2015 20:31:30 +0000 (13:31 -0700)]
perldelta: Fix typo

9 years agolib/B/Deparse.t: Suppress unescaped qr/{/ message
Karl Williamson [Wed, 18 Feb 2015 20:00:07 +0000 (13:00 -0700)]
lib/B/Deparse.t: Suppress unescaped qr/{/ message

9 years agoUnicode::UCD: Add charprops_all() public function
Karl Williamson [Wed, 18 Feb 2015 19:24:42 +0000 (12:24 -0700)]
Unicode::UCD: Add charprops_all() public function

9 years agoUnicode::UCD: Add charprop public function
Karl Williamson [Wed, 18 Feb 2015 19:16:00 +0000 (12:16 -0700)]
Unicode::UCD: Add charprop public function

9 years agolib/Unicode/UCD.t: White-space only
Karl Williamson [Sun, 15 Feb 2015 16:38:08 +0000 (09:38 -0700)]
lib/Unicode/UCD.t: White-space only

This aligns to a tab-stop, making it easier for me to type the next
commit

9 years agolib/Unicode/UCD.t: Convert some single quotes to double
Karl Williamson [Sat, 14 Feb 2015 16:29:55 +0000 (09:29 -0700)]
lib/Unicode/UCD.t: Convert some single quotes to double

This is in part because "" is easier to read than '', but more
importantly a future commit will want to have double quotish context in
places, and this makes it easier for me to convert to that.

9 years agoUnicode::UCD::prop_value_aliases() Don't return invalid value
Karl Williamson [Wed, 18 Feb 2015 18:57:02 +0000 (11:57 -0700)]
Unicode::UCD::prop_value_aliases() Don't return invalid value

Prior to this commit, if you said

    prop_value_aliases("Any", "foo")

it would return "foo".  But there really aren't any synonyms for the
"Any" property values, so it should return undef instead.

9 years agoUnicode::UCD: Pod corrections, clarifications
Karl Williamson [Sat, 14 Feb 2015 16:34:39 +0000 (09:34 -0700)]
Unicode::UCD: Pod corrections, clarifications

9 years ago[perl #123874] fix argument underflow for pack()
Hugo van der Sanden [Wed, 18 Feb 2015 14:59:53 +0000 (14:59 +0000)]
[perl #123874] fix argument underflow for pack()

NEXTFROM() modified the item count while testing it, so the next use
saw the count (of -1) as non-zero and ended up trying to write ~1 bytes.

9 years ago[perl #123870] fixup trie runtime debug output
Hugo van der Sanden [Wed, 18 Feb 2015 10:55:15 +0000 (10:55 +0000)]
[perl #123870] fixup trie runtime debug output

The trie_words av may not exist if -Dr was not enabled at compile time,
so we must be more defensive for debug output at runtime.

9 years agoRemove some remnants of x2p, which was removed by commit e5a8a0fbd7
Steve Hay [Wed, 18 Feb 2015 08:26:36 +0000 (08:26 +0000)]
Remove some remnants of x2p, which was removed by commit e5a8a0fbd7

References to PERL_FOR_X2P probably have to stay in case any third-party
code is #defining it. Likewise, NO_PERL_TYPEDEFS in vms/vmsish.h has to
stay because it actually is used by App::a2p.

Cross/Makefile.SH.patch still needs regenerating since it contains
references to x2p in the top-level Makefile.SH which were removed by
e5a8a0fbd7.

A few other references to x2p/a2p/find2p/s2p still remain, which I am less
sure about removing, plus some others in documentation which should stay
anyway.

9 years agoMake nowait spawn return correct pid on VMS.
Craig A. Berry [Wed, 18 Feb 2015 03:27:42 +0000 (21:27 -0600)]
Make nowait spawn return correct pid on VMS.

Apparently I missed this back in eed5d6a149b02.  When spawning a
subprocess without waiting, the return value is the pid of that
process, but we need to let pp_system know we're doing that so it
doesn't manipulate the value as if it were an exit status.

One symptom was that t/test.pl's watchdog would sometimes report
that it had failed to start when in fact it had started just fine.

9 years ago[perl #123852] avoid capture side-effects under noncapture flag
Hugo van der Sanden [Tue, 17 Feb 2015 03:27:33 +0000 (03:27 +0000)]
[perl #123852] avoid capture side-effects under noncapture flag

//n was implemented by avoiding the primary side-effects of compiling
a capture when the flag was turned on; however some secondary effects
still occurred later in the same function, by using the value of the
'paren' variable - even as far as causing coredumps.

Setting paren to ':' when NOCAPTURE is enabled makes the rest of the
function act just as if it had parsed (?:...) instead of (...).

9 years agoFix where we look for the Sun Studio compiler
Chris 'BinGOs' Williams [Tue, 17 Feb 2015 18:01:20 +0000 (18:01 +0000)]
Fix where we look for the Sun Studio compiler

It can be solarisstudio* as well

9 years agopp_pack.c: White-space only
Karl Williamson [Fri, 12 Dec 2014 00:21:45 +0000 (17:21 -0700)]
pp_pack.c: White-space only

This outdents some code whose enclosing block was removed in the
previous commit

9 years agopp_pack.c: Rmv useless code
Karl Williamson [Fri, 12 Dec 2014 00:17:45 +0000 (17:17 -0700)]
pp_pack.c: Rmv useless code

As noted in the thread starting at
http://nntp.perl.org/group/perl.perl5.porters/223366
and in the comments added in this commit, strings packed in 'u' format
don't need any UTF-8ness special handling, so the code that did that can
be removed.

9 years agopp_pack.c: Refactor to remove #if EBCDIC
Karl Williamson [Wed, 10 Dec 2014 03:57:52 +0000 (20:57 -0700)]
pp_pack.c: Refactor to remove #if EBCDIC

This commit causes the same code to be executed whether on an ASCII or
EBCDIC platform.

9 years agopp_pack.c: Change name of some static functions
Karl Williamson [Wed, 10 Dec 2014 00:45:35 +0000 (17:45 -0700)]
pp_pack.c: Change name of some static functions

Early code tends to conflate the terms Unicode and UTF-8.  I find that
confusing.

9 years agoperldsc: Wrap too-long verbatim lines
Karl Williamson [Tue, 17 Feb 2015 16:42:56 +0000 (09:42 -0700)]
perldsc: Wrap too-long verbatim lines

9 years agot/io/bom.t: Generalize for non-ASCII platforms
Karl Williamson [Sat, 6 Apr 2013 16:15:05 +0000 (10:15 -0600)]
t/io/bom.t: Generalize for non-ASCII platforms

This commit was updated on February 17, 2015 before being pushed to
blead to incorporate work from  Hugo van der Sanden and Dagfinn Ilmari
Mannsåker

9 years agoUpdate Unicode-Collate to CPAN version 1.11
Chris 'BinGOs' Williams [Tue, 17 Feb 2015 14:49:50 +0000 (14:49 +0000)]
Update Unicode-Collate to CPAN version 1.11

  [DELTA]

1.11  Tue Feb 17 21:23:03 2015
    - XS: [rt.cpan.org #102024] remove extranous const casting

9 years agofix handling of non-strict \x{}
Hugo van der Sanden [Mon, 16 Feb 2015 02:42:02 +0000 (02:42 +0000)]
fix handling of non-strict \x{}

It should be equivalent to \x{0}.

9 years agoRemove dead function from vms/vms.c
Craig A. Berry [Tue, 17 Feb 2015 01:14:30 +0000 (19:14 -0600)]
Remove dead function from vms/vms.c

fix_bare_dirnames is not used anywhere and doesn't actually do
anything.  It appears to be a placeholder for something that has
not materialized in the 10 years of it existence, so time to tidy
up.

9 years agoperldelta for 816b056ffb99, 3bea78d24634
Tony Cook [Tue, 17 Feb 2015 00:00:37 +0000 (11:00 +1100)]
perldelta for 816b056ffb993bea78d24634

9 years agoh2ph: correct handling of hex constants for the preamble
Tony Cook [Mon, 16 Feb 2015 04:57:00 +0000 (15:57 +1100)]
h2ph: correct handling of hex constants for the preamble

Previously they were treated as identifiers resulting in code
generated like C< &0xFFF >.

We also try to prevent compile-time warnings from large hex integers,
the user isn't responsible for the generated code, so we delay those
warnings to run-time.

9 years agoFix Errno.pm generation for gcc-5.0
Petr Písař [Wed, 11 Feb 2015 14:46:37 +0000 (15:46 +0100)]
Fix Errno.pm generation for gcc-5.0

gcc-5.0 -E interleaves now line numbers with expended macros, so that
the generated errno.c will be preprocessed to

EBFONT => [[
           59
                ]]

which is hard to parse in in line-based reader.

So use -P option with gcc >= 5.0. Global -P usage would break makedepend,
global -ftrack-macro-expansion=0 would break lib/h2ph.t.

RT#123784

9 years agoperldelta for 436b3d12366a, 2631bbca12a9
Tony Cook [Mon, 16 Feb 2015 22:59:58 +0000 (09:59 +1100)]
perldelta for 436b3d12366a2631bbca12a9

9 years ago[perl #123836] dump LABEL shouldn't be goto LABEL
Tony Cook [Mon, 16 Feb 2015 00:55:24 +0000 (11:55 +1100)]
[perl #123836] dump LABEL shouldn't be goto LABEL

9 years ago[perl #123836] basic tests for dump, catches the issue as a TODO
Tony Cook [Mon, 16 Feb 2015 00:54:25 +0000 (11:54 +1100)]
[perl #123836] basic tests for dump, catches the issue as a TODO

9 years agoModernize the code - "my" declarations and comments.
Shlomi Fish [Mon, 16 Feb 2015 14:46:29 +0000 (16:46 +0200)]
Modernize the code - "my" declarations and comments.

For: RT #123636

9 years agoSocket updated to CPAN version 2.018
Chris 'BinGOs' Williams [Mon, 16 Feb 2015 14:53:54 +0000 (14:53 +0000)]
Socket updated to CPAN version 2.018

  [DELTA]

2.018   2015/02/12 13:42:41
        [BUGFIXES]
         * Fix for "addr is not a string" test to use SvPOKp() before 5.18

9 years agoCPAN has version 5.20150214 now
Chris 'BinGOs' Williams [Mon, 16 Feb 2015 14:52:56 +0000 (14:52 +0000)]
CPAN has version 5.20150214 now

9 years agoDeparse state sub defined in inner subs
Father Chrysostomos [Wed, 31 Dec 2014 00:32:59 +0000 (16:32 -0800)]
Deparse state sub defined in inner subs

When we have ‘state sub x; sub { sub x { ... } }’, the actual &x sub-
routine  is not contained in the pad of the anonymous sub.  There is
only a pad entry pointing to the outer pad.  So we have go digging for
it by following ->OUTSIDE pointers.

This is only a partial fix.  If you wrap the code above in eval
'sub{...}' and deparse it, the state sub definition is missing,
because it is not available via ->OUTSIDE pointers.  I do not know
yet whether it is possible to get to it.  The to-do test in Deparse.t
used to croak.

9 years agoregen uconfig
H.Merijn Brand [Mon, 16 Feb 2015 09:38:48 +0000 (10:38 +0100)]
regen uconfig

9 years agoSmall doc change
H.Merijn Brand [Mon, 16 Feb 2015 09:08:28 +0000 (10:08 +0100)]
Small doc change

9 years agoRMG: Encourage use of http://analysis.cpantesters.org/beforemaintrelease
Steve Hay [Mon, 16 Feb 2015 08:50:00 +0000 (08:50 +0000)]
RMG: Encourage use of analysis.cpantesters.org/beforemaintrelease

9 years agorestore VC with dmake after commit "Fix minitest (op/taint.t) for GCC 4.x"
Daniel Dragan [Mon, 16 Feb 2015 05:03:07 +0000 (00:03 -0500)]
restore VC with dmake after commit "Fix minitest (op/taint.t) for GCC 4.x"

commit dbf36b7ad7 broke VC with dmake building by not defining the
test-prep-gcc target for VC builds. Instead of using $(NOOP), which incurs
the cost of dmake lauching another cmd.exe process to execute "rem", use
make lang preprocessor directives. Example of error

C:\perl521\src\win32>dmake test
.......normal building here.....
........cut.........
..\miniperl.exe -I..\lib ..\pod\perlmodlib.PL -q ..
..\perl.exe -f ..\pod\buildtoc -q
dmake:  Error: -- `test-prep-gcc' not found, and can't be made

C:\perl521\src\win32>