Father Chrysostomos [Sat, 27 Aug 2011 12:46:25 +0000 (05:46 -0700)]
Testing deparsing of CORE::not
Craig A. Berry [Sat, 27 Aug 2011 22:26:25 +0000 (17:26 -0500)]
Get coreamp.t passing on VMS.
The last record in a file ends with a newline willy nilly, so we
might as well write (and test for) one explicitly.
The name of a directory file ends with .DIR, so we need to allow
for that.
File::Temp panics when trying to delete the directory that is the
current working directory, so save where we were and restore to it
before clean-up.
Karl Williamson [Mon, 22 Aug 2011 15:26:09 +0000 (09:26 -0600)]
pp.c: Use built-in case tables for ords < 256
Previously, all case changing on utf8-encoded strings used the tables on
disk, under the off-chance that there was a user-defined case change
override in effect. Now that that feature has been removed, this can't
happen, so we can use the existing built-in tables.
This code has been present and ifdef'd out since 5.10.1. New compiler
warnings forced a few other changes besides removing the #if statements
Running some primitive benchmarks showed that this sped up upper-casing of
utf8 strings in the latin1 range by 2 orders of magnitude.
Karl Williamson [Mon, 22 Aug 2011 15:25:11 +0000 (09:25 -0600)]
pp.c: Change comment
This now reflects Tom Christiansen's and my current thinking about Greek
Final Sigma
Florian Ragwitz [Thu, 25 Aug 2011 15:50:26 +0000 (17:50 +0200)]
Make pod2html a regular script without substitutions
This will make the CPAN dist easier. For the perl core, we still need
substitutions to get the right she-bang as we don't go through EU::MM to fix it
for us. For that, we add utils/pod2html.PL.
Florian Ragwitz [Thu, 25 Aug 2011 15:00:07 +0000 (17:00 +0200)]
Give Pod-Html a more modern dist layout
Father Chrysostomos [Sat, 27 Aug 2011 06:50:06 +0000 (23:50 -0700)]
&CORE::reset()
This commit allows &CORE::reset to be called through references and
via ampersand syntax. pp_reset is modified to take into account the
nulls pushed on to the stack in pp_coreargs, which happens because
pp_coreargs has no other way to tell reset how many arguments it’s
actually getting. See commit
0163043a for details.
Father Chrysostomos [Sat, 27 Aug 2011 06:31:36 +0000 (23:31 -0700)]
GVs of localised arrays and hashes should be refcounted
Otherwise the GV can be freed before the scope-popping code can put
the old entry back in it:
$ perl -le 'local @{"x"}; delete $::{x}'
Bus error
$ perl -le 'local %{"x"}; delete $::{x}'
Bus error
Father Chrysostomos [Sat, 27 Aug 2011 05:28:52 +0000 (22:28 -0700)]
&CORE::foo() for (sys)read and recv
These are grouped together because they all have \$ in their
prototypes.
This commit allows the subs in the CORE package under those names to
be called through references and via &ersand syntax.
The coreargs op in the subroutine is marked with the OPpSCALARMOD
flag. (scalar_mod_type in op.c returns true for these three ops,
indicating that the OA_SCALARREF parameter is \$, not \[$@%(&)*].)
pp_coreargs uses that flag to decide what arguments to reject.
Father Chrysostomos [Sat, 27 Aug 2011 05:05:40 +0000 (22:05 -0700)]
Add OPpCOREARGS_SCALARMOD flag
pp_coreargs will use this to distinguish between the \$ and \[$@%*]
prototypes.
Father Chrysostomos [Sat, 27 Aug 2011 04:57:59 +0000 (21:57 -0700)]
Update the comments at the top of t/op/core*.t
Father Chrysostomos [Sat, 27 Aug 2011 04:48:47 +0000 (21:48 -0700)]
Rename t/op/core*.t
Originally, coresubs.t was going to be for generic tests and
coreinline.t was going to be for inlining. But the latter ended
up testing other things than inlining, the former testing just
&ersand() calls. So this commits renames coresubs.t to coreamp.t
and coreinline.t to coresubs.t.
Father Chrysostomos [Fri, 26 Aug 2011 17:12:01 +0000 (10:12 -0700)]
&CORE::rand()
This commit allows &CORE::rand to be called through references and
via ampersand syntax. pp_rand is modified to take into account the
nulls pushed on to the stack in pp_coreargs, which happens because
pp_coreargs has no other way to tell rand how many arguments it’s
actually getting. See commit
0163043a for details.
Father Chrysostomos [Sat, 27 Aug 2011 01:19:14 +0000 (18:19 -0700)]
&CORE::open()
This commit allows &CORE::open to be called through references or with
ampersand syntax. It modifies pp_coreargs not to push nulls for ops
that require a pushmark.
Father Chrysostomos [Fri, 26 Aug 2011 21:52:06 +0000 (14:52 -0700)]
Reverse the order of two tests in gv.t
so that the first does not make the second always pass.
The bug that
99fc7eca4 fixed would never had occurred if
they had been that way to begin with.
Father Chrysostomos [Thu, 25 Aug 2011 16:52:13 +0000 (09:52 -0700)]
&CORE::mkdir()
This commit allows &CORE::mkdir to be called through references and
via ampersand syntax. pp_mkdir is modified to take into account the
nulls pushed on to the stack in pp_coreargs, which happens because
pp_coreargs has no other way to tell mkdir how many arguments it’s
actually getting.
Father Chrysostomos [Wed, 24 Aug 2011 16:09:58 +0000 (09:09 -0700)]
&CORE::lock()
This commit allows &CORE::lock to be called through references and
via ampersand syntax. It adds code to pp_coreargs for handling the
OA_SCALARREF case, though what it adds is currently lock-specific.
(Subsequent commits will address that.) Since lock returns the scalar
passed to it, not a copy, &CORE::lock needs to use op_leavesublv,
rather than op_leavesub. But it can’t be an lvalue sub, as
&CORE::lock = 3 should be disallowed. So we use the sneaky trick of
turning on the lvalue flag before attaching the op tree to the sub
(which causes newATTRSUB to use op_leavesublv), and then turning it
off afterwards.
Father Chrysostomos [Fri, 26 Aug 2011 19:40:13 +0000 (12:40 -0700)]
&CORE::index() and &CORE::rindex()
This commit allows &CORE::index and &CORE::rindex to be called through
references and via ampersand syntax. pp_index is modified to take
into account the nulls pushed on to the stack in pp_coreargs, which
happens because pp_coreargs has no other way to tell pp_index how many
arguments it’s actually getting. See commit
0163043a for details.
Father Chrysostomos [Fri, 26 Aug 2011 17:00:13 +0000 (10:00 -0700)]
Update UNIVERSAL::VERSION docs following 9bf41c1
Father Chrysostomos [Fri, 26 Aug 2011 16:49:37 +0000 (09:49 -0700)]
&CORE::gmtime() and &CORE::localtime()
This commit allows &CORE::gmtime and &CORE::localtime to be called
through references and via ampersand syntax. pp_gmtime is modified
to take into account the nulls pushed on to the stack in pp_coreargs,
which happens because pp_coreargs has no other way to tell pp_gmtime
how many arguments it’s actually getting.
I was going to say ‘see commit
f6a1686942 for more details’, but found
out, to my horror, that most of the commit message was cut off. I
don’t know which commit-munging part of git is responsible, but I’ve
had similar problems with git am and git commit --amend. But, then,
this could have been sloppy copy and paste. Anyway, here is the
missing part:
Usually, an op that has optional arguments has the number of arguments
indicated with flags on the op itself:
$ ./perl -Ilib -MO=Concise -e 'binmode 1'
6 <@> leave[1 ref] vKP/REFC ->(end)
1 <0> enter ->2
2 <;> nextstate(main 1 -e:1) v:{ ->3
5 <@> binmode vK/1 ->6
- <0> ex-pushmark s ->3
4 <1> rv2gv sK*/1 ->5
3 <$> gv(*1) s ->4
-e syntax OK
$ ./perl -Ilib -MO=Concise -e 'binmode 1,2'
7 <@> leave[1 ref] vKP/REFC ->(end)
1 <0> enter ->2
2 <;> nextstate(main 1 -e:1) v:{ ->3
6 <@> binmode vK/2 ->7
- <0> ex-pushmark s ->3
4 <1> rv2gv sK*/1 ->5
3 <$> gv(*1) s ->4
5 <$> const(IV 2) s ->6
-e syntax OK
Notice the /1 vs /2 on the binmode op.
With a CORE sub, we have a single op for both cases. So, what this
commit does is set the number of arguments to the maximum, push nulls
on to the stack in pp_coreargs (actually, it was already set up to
do it, so there is no change there), and have the pp_ functions for
each op that has optional arguments do a null check after popping
the stack.
pp_binmode already does a null check, but other pp_ functions will
need to be modified. Since each one is different, those will come in
separate commits.
This is what &CORE::binmode’s op tree looks like:
$ ./perl -Ilib -MO=Concise,CORE::binmode -e 'BEGIN{\&CORE::binmode}'
CORE::binmode:
3 <1> leavesub[1 ref] K/REFC,1 ->(end)
2 <@> binmode sK/2 ->3
- <0> ex-pushmark s ->1
1 <$> coreargs(IV 212) s ->2
-e syntax OK
Father Chrysostomos [Fri, 26 Aug 2011 16:26:46 +0000 (09:26 -0700)]
perldelta for
f132ae69 (*{undef})
Father Chrysostomos [Fri, 26 Aug 2011 15:25:45 +0000 (08:25 -0700)]
Remove now-unnecessary len check from pp.c:S_rv2gv
This check, added by
ed996e63f6, is no longer necessary as of commit
f132ae694c, since PL_sv_undef takes a different path.
Father Chrysostomos [Fri, 26 Aug 2011 13:20:01 +0000 (06:20 -0700)]
Make *{undef} self-consistent
Commit
afd1915d made filehandle vivification work on hash and array
elements, but in doing so it accidentally changed
*{;undef} = 3;
to do the same thing as
*{""} = 3;
while leaving
*{$some_undefined_variable}
an error.
This commit adjusts the if() conditions in S_rv2gv (formerly in
pp_rv2gv) in pp.c to make PL_sv_undef follow the same path as before.
It also removes the uninit tests from lib/warnings/pp, since they
are now errors. The uninit warning in rv2gv is only triggered now
when it is implicit, as in close(). That is already tested in
lib/warnings/9uninit.
Father Chrysostomos [Fri, 26 Aug 2011 12:55:34 +0000 (05:55 -0700)]
perlop: Minor consistency tweak
Make the indentation in this example match the surrounding
examples.
Father Chrysostomos [Fri, 26 Aug 2011 12:54:58 +0000 (05:54 -0700)]
[perl #93358] Clarify => quoting
The perlop manpage was stating ‘the left operand’, which was
not entirely correct, as ‘time.shift =>’ quotes just the shift,
not the time (nor does it see the whole as not being an ident-
ifier and refuse to quote anything).
Florian Ragwitz [Fri, 26 Aug 2011 13:26:38 +0000 (15:26 +0200)]
Re-generate uconfig.h
This makes porting/regen.t pass again.
Florian Ragwitz [Fri, 26 Aug 2011 12:35:51 +0000 (14:35 +0200)]
Bump the perlfaq version for a new cpan release
H.Merijn Brand [Fri, 26 Aug 2011 11:02:23 +0000 (13:02 +0200)]
Remove unwanted space in comments
H.Merijn Brand [Fri, 26 Aug 2011 10:14:53 +0000 (12:14 +0200)]
Revert "Fix Configure's csym test for gcc's link time optimisation"
There are too many related problems arising from this change
This reverts commit
e820c6d6a6d0a8828aa68a6895696b659c471f2f.
Father Chrysostomos [Fri, 26 Aug 2011 04:36:12 +0000 (21:36 -0700)]
Correct mro_get_linear_isa’s docs
This has been out of date since the MRO plugin API was added
(5.10.1, I think).
Father Chrysostomos [Fri, 26 Aug 2011 01:12:35 +0000 (18:12 -0700)]
&CORE::getpgrp()
This commit allows &CORE::getpgrp to be called through references and
via ampersand syntax. pp_getpgrp is modified to take into account the
nulls pushed on to the stack in pp_coreargs, since pp_coreargs has no
other way to tell getpgrp how many arguments it’s actually getting.
Father Chrysostomos [Fri, 26 Aug 2011 01:10:53 +0000 (18:10 -0700)]
&CORE::exit()
This commit allows &CORE::exit to be called through references and
via ampersand syntax. pp_exit is modified to take into account the
nulls pushed on to the stack in pp_coreargs, since pp_coreargs has no
other way to tell exit how many arguments it’s actually getting.
Father Chrysostomos [Fri, 26 Aug 2011 03:02:58 +0000 (20:02 -0700)]
pp.c: Suppress stupid compiler warning
Father Chrysostomos [Fri, 26 Aug 2011 01:06:23 +0000 (18:06 -0700)]
&CORE::foo() for dbmopen and dbmclose
This commit allows the subs in the CORE package for close, getc and
readline to be called through references and via ampersand syntax. A
special case for each of them is added to pp_coreargs to deal with
calls with no arguments. Pushing a null on to the stack (which I’m
doing for other ops) won’t work, as a null already means something for
these cases: close($f) won’t vivify a typeglob if $f is a string, so
the implicit rv2gv pushes a null on to the stack.
Father Chrysostomos [Fri, 26 Aug 2011 00:59:12 +0000 (17:59 -0700)]
&CORE::foo() for close, getc and readline
This commit allows the subs in the CORE package for close, getc and
readline to be called through references and via ampersand syntax. The
pp functions are modified to take into account the nulls that coreargs
pushes on to the stack to indicate that there is no argument.
Craig A. Berry [Thu, 25 Aug 2011 22:16:13 +0000 (17:16 -0500)]
void rather than empty parameter for Perl___notused.
Without this, the VMS compiler, with warnings cranked up to level 4,
emits pages and pages of things like:
dVAR;
....^
%CC-I-NOPARMLIST, The declaration of the function Perl___notused has
an empty parameter list. If the function has parameters, they should
be declared here; if it has no parameters, "void" should be specified
in the parameter list.
at line number 100 in file MDA0:[SMOKE.blead]perl.c;1
Over 2,000 of these plus other warnings yields a smoke report of 750K,
which is quite a bit over the 400K limit of the perl.org mailing lists,
not to mention being a slow read.
Father Chrysostomos [Thu, 25 Aug 2011 21:46:55 +0000 (14:46 -0700)]
[perl #92376] Change - to : in time template
Father Chrysostomos [Thu, 25 Aug 2011 21:33:03 +0000 (14:33 -0700)]
&CORE::foo() for @ and $@ prototypes, except unlink
This commit allows the CORE subroutines for functions with @
and $@ prototypes to be called through references and via amper-
sand syntax.
unlink is not included in this commit, as it requires special casing
due to its use of implicit $_.
Since these functions require a pushmark, and since it has to come
between two things that pp_coreargs does, it’s easiest to flag the
coreargs op (with the OPpCOREARGS_PUSHMARK flag added in the previous
commit) and call pp_pushmark directly from pp_coreargs.
Father Chrysostomos [Mon, 22 Aug 2011 00:27:58 +0000 (17:27 -0700)]
Add OPpCOREARGS_PUSHMARK flag
This will be used to tell pp_coreargs when it needs to call
pp_pushmark.
For those functions that need a pushmark, it has to come between
two things that pp_coreargs does; so the easiest way is to use
this flag.
Father Chrysostomos [Sun, 21 Aug 2011 18:59:44 +0000 (11:59 -0700)]
&CORE::caller()
This commit allows &CORE::caller to be called through references and
via ampersand syntax. pp_caller is modified to take into account
two things:
1) pp_coreargs pushes a null on to the stack, since it has no other
way to tell caller whether it has an argument.
2) The value coming from pp_coreargs (when not null) is off by
one. The OPpOFFYBONE flag was added in commit
93f0bc4935 for
this purpose.
pp_coreargs is also modified, since it assumed till now that an
optional first argument was an implicit $_.
Father Chrysostomos [Sun, 21 Aug 2011 17:57:34 +0000 (10:57 -0700)]
&CORE::bless()
This commit allows &CORE::bless to be called through references and
via ampersand syntax. pp_bless is modified to take into account the
nulls pushed on to the stack in pp_coreargs, since pp_coreargs has no
other way to tell bless how many arguments it’s actually getting.
Father Chrysostomos [Thu, 25 Aug 2011 19:43:54 +0000 (12:43 -0700)]
&CORE::binmode()
This commit allows &CORE::binmode to be called through references and
via ampersand syntax.
Usually, an op that has optional arguments has the number of arguments
indicated with flags on the op itself:
Father Chrysostomos [Thu, 25 Aug 2011 19:38:15 +0000 (12:38 -0700)]
[perl #93320] localising @DB::args leads to coredump
This script, from the RT ticket, crashes:
#!/usr/bin/perl
sub cl{
package DB;
@DB::args = ();
return caller(shift);
}
sub f{
local @DB::args;
my @z = cl($_) for (1..3);
}
f(1,2,3); f(1,2,3);
__END__
PL_dbargs is not refcounted, and it’s not set until pp_caller first
tries to write to it. If that happens when @DB::args is localised,
then the array will be freed on scope exit, leaving PL_dbargs pointing
to a freed SV.
This crash can be reproduced more simply this way:
sub {
package DB;
()=caller(0);
undef *DB::args;
()=caller(0);
}->();
So, basically, pp_caller has to re-fetch PL_dbargs from the %DB::
stash each time it sets it. It cannot rely on the cached value.
(So now I’m wondering whether we even need PL_dbargs.)
Father Chrysostomos [Thu, 25 Aug 2011 16:50:02 +0000 (09:50 -0700)]
Allow ampersand calls for CORE subs with $*$$**$ protos
This enables ampersand calls and calls through references for CORE
subs that have * and $ in their prototypes and a fixed number of
arguments.
Usually, the *-prototyped ops have their child ops wrapped in rv2gv’s
(*{}) implicitly. The rv2gv op is sometimes flagged as an autoviv-
ificatory op, such as the first argument to accept() or open().
S_is_handle_constructor contains the list of ops that turn on
that flag.
This commit makes the coreargs op use a couple of flags to serve the
same purpose. pp_coreargs itself calls S_rv2gv (split out from
pp_rv2gv recently for precisely this purpose) with arguments based on
its own flags.
Currently the autovivified glob gets a name like main::_GEN_0 instead
of main::$a. I think we can live with that.
Father Chrysostomos [Sat, 20 Aug 2011 06:04:04 +0000 (23:04 -0700)]
Add private coreargs flags for vivifying GVs
Father Chrysostomos [Sat, 20 Aug 2011 01:11:23 +0000 (18:11 -0700)]
Move most of pp_rv2gv into a static function
This allows the glob-deref logic to be use by other functions in pp.c,
like pp_coreargs.
Father Chrysostomos [Thu, 25 Aug 2011 16:36:29 +0000 (09:36 -0700)]
Add Frederic Briere to AUTHORS
Frederic Briere [Thu, 25 Aug 2011 16:35:26 +0000 (09:35 -0700)]
Attribute::Handlers: "my" should be "till"
Father Chrysostomos [Thu, 25 Aug 2011 16:33:48 +0000 (09:33 -0700)]
Attribute::Handlers: correct spelling
Father Chrysostomos [Thu, 25 Aug 2011 16:25:21 +0000 (09:25 -0700)]
Dumper.xs: Suppress compiler warning
Father Chrysostomos [Thu, 25 Aug 2011 15:43:22 +0000 (08:43 -0700)]
coresubs.t: Minor clean-up
This fixes some infelicities in the new tests I added to this
file recently.
Father Chrysostomos [Fri, 19 Aug 2011 17:57:41 +0000 (10:57 -0700)]
Invert the list of &-able functions in gv.c
The list of those that do not support &CORE::foo() syntax is now
shorter than the list of those that do. In subsequent commits
it will get even shorter.
Father Chrysostomos [Thu, 25 Aug 2011 15:34:15 +0000 (08:34 -0700)]
Make sure coresubs.t tests all &-able funcs
Leo Lapworth [Thu, 25 Aug 2011 10:08:23 +0000 (11:08 +0100)]
replace old bookmarks.cpan link with perl.org and learn.perl.org
Nicholas Clark [Thu, 25 Aug 2011 10:06:53 +0000 (12:06 +0200)]
Merge the refactored makedef.pl into blead.
This eliminates global.sym
Nicholas Clark [Thu, 25 Aug 2011 10:04:23 +0000 (12:04 +0200)]
Note the demise of global.sym in perldelta.
Nicholas Clark [Sun, 21 Aug 2011 15:50:26 +0000 (17:50 +0200)]
In embed.pl, inline walk_table() into its only caller.
Nicholas Clark [Sun, 21 Aug 2011 15:37:41 +0000 (17:37 +0200)]
Eliminate global.sym, as makedef.pl can generate it internally.
global.sym was a file listing the exported symbols, generated by regen/embed.pl
from embed.fnc and regen/opcodes, which was only used by makedef.pl
Move the code that generates global.sym from regen/embed.pl to makedef.pl,
and thereby eliminate the need to ship a 907 line generated file.
Nicholas Clark [Sun, 21 Aug 2011 14:48:51 +0000 (16:48 +0200)]
Add regen/embed_lib.pl, for the code that processes embed.fnc and regen/opcodes
Move setup_embed() and the helper functions add_level() and current_group()
to it from regen/embed.pl
Nicholas Clark [Sun, 21 Aug 2011 14:27:15 +0000 (16:27 +0200)]
In embed.pl, move processing embed.fnc and regen/opcodes into a function.
Nicholas Clark [Sun, 21 Aug 2011 13:47:56 +0000 (15:47 +0200)]
In embed.pl, refactor the handling of varargs *_nocontext wrappers.
This is what commit
125218eb5a6d12e7 should have been. :-)
Nicholas Clark [Sun, 21 Aug 2011 12:58:21 +0000 (14:58 +0200)]
In makedef.pl, move handling of $ARGS{TARG_DIR} to the open statements.
This simplifies the code, and eliminates 5 lexicals used solely to hold
filename constants.
Nicholas Clark [Sun, 21 Aug 2011 12:30:14 +0000 (14:30 +0200)]
In makedef.pl, defaulting $ARGS{TARG_DIR} to '' simplifies the code.
Nicholas Clark [Sun, 21 Aug 2011 12:06:33 +0000 (14:06 +0200)]
Test the TARG_DIR argument to makedef.pl too.
Nicholas Clark [Sun, 21 Aug 2011 11:40:33 +0000 (13:40 +0200)]
In makedef.pl, no need to store 'config.h' in a variable.
Unlike the other files opened by makedef.pl, config.h is always opened in the
current directory, so the filename doesn't modifying if $ARGS{TARG_DIR} is set.
Nicholas Clark [Sun, 21 Aug 2011 11:29:26 +0000 (13:29 +0200)]
In embed.pl, simplify the code that parses regen/opcodes.
As @embed is sorted later, the order that we add entries to it doesn't matter.
Hence add them immediately, avoiding the need to iterate over the "seen" hash.
Nicholas Clark [Sun, 21 Aug 2011 11:09:46 +0000 (13:09 +0200)]
embed.pl was relying on embed.fnc starting with a pre-processor directive.
Fix this subtle bug, and add comments.
Steffen Mueller [Thu, 25 Aug 2011 06:19:40 +0000 (08:19 +0200)]
ExtUtils::ParseXS: Accept overridden input typemaps
This restores the ability to have code like this:
ret_type*
foo(bar, baz)
int bar
ThereIsNoTypemapForThisType* baz = somefunc($arg);
CODE:
...
Looks strange and indeed, it is. But it's documented in perlxs
to work, so we can't get away with breaking it. The heuristics
for determining whether to allow this use case is checking for
ST\( or \$arg being found in the initialization. That should take
care of all valid use of this feature and still die horribly in
cases where this is used by mistake (instead of a PREINIT block, etc).
Steffen Mueller [Sun, 21 Aug 2011 20:07:50 +0000 (22:07 +0200)]
ExtUtils::ParseXS: Explicitly require current version of submodules
Since there have been certain problems with parts of ExtUtils::ParseXS
being shadowed by older installations of the module, this commit adds
an explicit $VERSION to all submodules and requires them to have the
same version as the main module. This doesn't actually fix any problem,
but makes them more apparent as early as possible instead of failing
with obscure compile errors when bad C is generated from the original
XS.
Father Chrysostomos [Thu, 25 Aug 2011 07:23:48 +0000 (00:23 -0700)]
perldelat update
Father Chrysostomos [Thu, 25 Aug 2011 06:55:46 +0000 (23:55 -0700)]
perldelta: Move an entry under Testing
Father Chrysostomos [Fri, 19 Aug 2011 15:27:14 +0000 (08:27 -0700)]
Enable ampersand calls to CORE subs with $$$ prototypes
This applies to functions that just take plain scalar arguments, all
of which are mandatory. Functions that take optional arguments are
not supported yet. truncate() is not supported yet, either (its $$
prototype is not entirely veracious).
This commit enables those functions to be called via &CORE::foo() syn-
tax or through references.
You can now encrypt a string like this: "string"->CORE::crypt($salt).
Each function’s op tree is like this:
$ ./perl -Ilib -MO=Concise,CORE::atan2 -e 'BEGIN{\&CORE::atan2}'
CORE::atan2:
3 <1> leavesub[1 ref] K/REFC,1 ->(end)
2 <@> atan2[t1] sK ->3
- <0> ex-pushmark s ->1
1 <$> coreargs(IV 100) s ->2
-e syntax OK
This commit adds code to ck_fun to skip the argument check if
coresubs is present. Otherwise we get a ‘Not enough arguments for
atan2’ error.
Father Chrysostomos [Fri, 19 Aug 2011 05:19:25 +0000 (22:19 -0700)]
Simplify the CORE::__FOO__ op-generation code
It just happens that the (caller)[...] offsets for file and line are
the same as the keyword numbers. KEY___PACKAGE__ is 3 and (caller)[0]
returns the package, so keyword_number % 3 can be used for the offset
instead of an unwieldy switch block.
Father Chrysostomos [Fri, 19 Aug 2011 05:09:17 +0000 (22:09 -0700)]
Move coresub op-creation from gv.c to op.c
For functions that take handles as arguments, this code will need to
call static functions in op.c, like is_handle_constructor.
While we could make is_handle_constructor into a non-static function
and call it from gv.c, that seems backwards, as it would result in a
lot of op-manipulation code in the middle of gv.c.
So this commit creates a new function in op.c, called coresub_op,
which is only called from gv.c, from the &CORE::sub code.
Father Chrysostomos [Thu, 25 Aug 2011 06:31:28 +0000 (23:31 -0700)]
Allow ampersand calls to CORE subs with (_) proto
This commit adds all subs with a (_) prototype to the list of those
that can be called with ampersand syntax or through references. These
have bodies like this:
$ ./perl -Ilib -MO=Concise,CORE::abs -e 'BEGIN{\&CORE::abs}'
CORE::abs:
3 <1> leavesub[1 ref] K/REFC,1 ->(end)
2 <1> abs[t1] sK/1 ->3
1 <$> coreargs(IV 111) s ->2
-e syntax OK
coreargs fetches the caller’s $_ if there are no arguments passed.
Father Chrysostomos [Thu, 18 Aug 2011 22:55:59 +0000 (15:55 -0700)]
Add find_rundefsv2 function
Subs in the CORE package with a (_) prototype will use this.
This accepts a CV and a sequence number, so that one can
use it to find the $_ in the caller’s scope. It only uses
the topmost call of a subroutine that is being called recur-
sively, so it’s not really a general-purpose function. But
it suffices for &CORE::abs and friends.
Father Chrysostomos [Thu, 25 Aug 2011 05:39:28 +0000 (22:39 -0700)]
Test CORE::break’s prototype (just that this time)
Gerard Goossen [Tue, 16 Aug 2011 07:22:14 +0000 (09:22 +0200)]
Move making inplace sort and reverse away from the peephole optimiser to scalarvoid.
Why: The in place assignment is not just an optimisation but has
significant different behaviour and thus doesn't belong in the
peephole optimiser. Also the optree changes are unified and simpler.
Father Chrysostomos [Thu, 25 Aug 2011 06:15:17 +0000 (23:15 -0700)]
Revert "Test CORE::break’s prototype"
This reverts commit
e52d58aa5bea245b66786b4c9029e849a2be69d3.
I don’t quite know how I managed it, but I really screw up
this time! Two completely unrelated commits ended up getting
merged into one, so, to avoid confusion down the road, I’m
reverting it, only to reapply it shortly....
Father Chrysostomos [Thu, 25 Aug 2011 05:48:26 +0000 (22:48 -0700)]
&CORE::break
This is another nullary function that I forgot to give a body
to make it callable via &break syntax.
Father Chrysostomos [Thu, 25 Aug 2011 05:39:28 +0000 (22:39 -0700)]
Test CORE::break’s prototype
Father Chrysostomos [Thu, 25 Aug 2011 05:16:07 +0000 (22:16 -0700)]
[perl #71154] undef &$coderef consistency
$ perl -le' undef &{$x=sub{}}; $x->()'
Not a CODE reference at -e line 1.
undeffing an anonymous subroutine used to turn the ANON flag off,
causing it to bypass the condition apparently written for this situa-
tion in pp_entersub.
This commit stops cv_undef from turning off that flag, so now we get:
$ ./perl -le' undef &{$x=sub{}}; $x->()'
Undefined subroutine called at -e line 1.
Eric Brine [Thu, 25 Aug 2011 03:10:43 +0000 (20:10 -0700)]
Trim dead code in do_kv.
A small piece of code in do_kv has three bugs:
- TARG could have been returned as an lvalue, so its refcount
could be greater than 1, resulting in data getting clobbered.
(See RT#20933 for previously fixed occurrence of this bug).
- LvTARG is refcounted, so it's buggy to just NULL it.
- TARG is returned without being initialised.
The first two bugs disappeared recently when we stopped putting
the lvalues in TARG for that op. The third remains.
However, it seems that code is never called. It can only be
called by putting NULL (not undef) on the Perl stack. I don't
see how that's possible here. The test suite never reaches that
code, so it seems it's just dead code.
Leo Lapworth [Wed, 24 Aug 2011 13:26:55 +0000 (14:26 +0100)]
Remove nvi (site no loger works) and put vim above Vile
Leo Lapworth [Wed, 24 Aug 2011 13:24:21 +0000 (14:24 +0100)]
Remove Elvis editor as link broken
Leo Lapworth [Wed, 24 Aug 2011 13:22:49 +0000 (14:22 +0100)]
Fix broken link
Leo Lapworth [Wed, 24 Aug 2011 13:18:21 +0000 (14:18 +0100)]
Remove a2ps as link does not work and even gnu.org uses the same link
Father Chrysostomos [Thu, 25 Aug 2011 01:04:26 +0000 (18:04 -0700)]
Make $class->method work when $class is tied
This little script:
sub TIESCALAR{bless[]}
sub FETCH{warn "fetching"; "main"}
sub bolgy { warn 'bolgy' }
tie my $a, "";
$a->bolgy;
Gives these outputs with various versions of perl:
$ pbpaste|perl5.6.2
fetching at - line 2.
fetching at - line 2.
bolgy at - line 3.
$ pbpaste|perl5.8.8
fetching at - line 2.
fetching at - line 2.
fetching at - line 2.
Can't call method "bolgy" without a package or object reference at - line 5.
$ pbpaste|perl5.8.9
fetching at - line 2.
fetching at - line 2.
fetching at - line 2.
fetching at - line 2.
bolgy at - line 3.
$ pbpaste|perl5.10.0
fetching at - line 2.
fetching at - line 2.
fetching at - line 2.
fetching at - line 2.
Can't call method "bolgy" without a package or object reference at - line 5.
$ pbpaste|perl5.10.1 # also 5.12.x
fetching at - line 2.
fetching at - line 2.
fetching at - line 2.
fetching at - line 2.
bolgy at - line 3.
$ pbpaste|perl5.14.0
fetching at - line 2.
fetching at - line 2.
fetching at - line 2.
fetching at - line 2.
fetching at - line 2.
fetching at - line 2.
Can't locate object method "bolgy" via package "main" (perhaps you forgot to load "main"?) at - line 5.
It’s worse than ever in 5.14.
What’s happening is that S_method_common is hanging on to the pointer
returned by SvPV, while continuing to call get-magic again and again.
So the pointer becomes invalid. I think it’s only by accident that
it worked in some versions.
This commit stops S_method_common from calling get-magic so many
times, solving both problems.
I’m afraid this conflicts with ongoing work to make method lookup
UTF8-clean, but I wanted to make a patch that could be backported.
Father Chrysostomos [Wed, 24 Aug 2011 21:33:31 +0000 (14:33 -0700)]
gmagic.t: Correct two test names
Father Chrysostomos [Wed, 24 Aug 2011 21:28:45 +0000 (14:28 -0700)]
Un-todo gmagic.t tests that now pass
Gerard Goossen [Wed, 24 Aug 2011 21:26:51 +0000 (14:26 -0700)]
[perl #97088] Prevent double get-magic in various cases
This patch prevents get-magic from executing twice during autovivifi-
cation when the op doing the autovivification is not directly nested
inside the dereferencing op.
This can happen in cases like this:
${ (), $a } = 1;
Previously (as of 5.13.something), the outer op was marked with the
OPpDEREFed flag, which indicated that get-magic had already been
called by the vivifying op (calling get-magic during vivification is
inevitable):
$ perl5.14.0 -MO=Concise -e '${ $a } = 1'
8 <@> leave[1 ref] vKP/REFC ->(end)
1 <0> enter ->2
2 <;> nextstate(main 2 -e:1) v:{ ->3
7 <2> sassign vKS/2 ->8
3 <$> const[IV 1] s ->4
6 <1> rv2sv sKRM*/DREFed,1 ->7 <-- right here
- <@> scope sK ->6
- <0> ex-nextstate v ->4
5 <1> rv2sv sKM/DREFSV,1 ->6
4 <#> gv[*a] s ->5
-e syntax OK
But in the ${()...} example above, there is a list op in the way that
prevents the flag from being set inside the peephole optimizer. It’s
not even possible to set it correctly in all cases, as in this exam-
ple, which would need it both set and not set depending on which
branch of the ternary operator is executed:
${ $x ? delete $a[0] : $a[0] } = 1
Instead of setting the OPpDEREFed flag, we now make a non-magic copy
of the SV in vivify_ref (the first time get-magic is executed).
Gerard Goossen [Wed, 24 Aug 2011 11:39:55 +0000 (13:39 +0200)]
Add tests for autovivication combined with get-magic, some of which are TODO.
Father Chrysostomos [Wed, 24 Aug 2011 18:10:22 +0000 (11:10 -0700)]
Add missing files from Archive::Extract 0.56
Nicholas Clark [Tue, 23 Aug 2011 12:10:58 +0000 (14:10 +0200)]
For microperl, prefix generateuudmap and the 3 generated headers with u.
Without this, the main Makefile and Makefile.micro interfere with each other,
as they both generate the same 5 files, and both think that they can delete
them with their respective clean targets.
Nicholas Clark [Tue, 23 Aug 2011 11:47:59 +0000 (13:47 +0200)]
Group 3 headers as $(generated_headers) in microperl's Makefile.
Commit
9387abf86b0530ac missed this.
Father Chrysostomos [Wed, 24 Aug 2011 01:15:48 +0000 (18:15 -0700)]
Call get-magic once for defined ${...}
This example:
sub TIESCALAR { bless[]}
sub FETCH { warn "fetching"; "\cTAINT" }
tie my $a, "";
defined $$a;
prints ‘fetching’ three times in 5.8.8, five times (!) in 5.10-12,
four times in 5.14, and three times in blead as of
ed996e63f6. Now it
only happens once.
It was commit
7a5fd60d4c that increased the number of fetches in 5.10,
but I haven’t checked which commits reduced it in 5.14.
Chip Salzenberg [Tue, 23 Aug 2011 23:51:28 +0000 (16:51 -0700)]
SVTYPEMASK must be cast to (svtype) when comparing to SvTYPE()
Father Chrysostomos [Tue, 23 Aug 2011 21:54:34 +0000 (14:54 -0700)]
Stop readline($foo) from autovivifying
Currently <$foo> does not autovivify, but readline($foo) does, due to
the way pp_readline calls pp_rv2gv directly, with PL_op still holding
a pp_readline op, whose flags may have completely different meanings.
readline uses the OPf_SPECIAL flag to distinguish <$foo> from readline
($foo). rv2gv uses it to determine whether to autovivify; hence the
discrepancy.
Father Chrysostomos [Tue, 23 Aug 2011 21:36:46 +0000 (14:36 -0700)]
close($undef) should not croak_no_modify
Commit
ac53db4c3f7e fixed bug #31767 (open $1 not dying), but put the
SvREADONLY check in the wrong spot, causing this bug:
$ perl -lwe 'no warnings "once"; close $x; close $+'
Name "main::x" used only once: possible typo at -e line 1.
Use of uninitialized value $x in ref-to-glob cast at -e line 1.
Modification of a read-only value attempted at -e line 1.
It shouldn’t be dying if I’m not trying to modifying it.
Father Chrysostomos [Tue, 23 Aug 2011 21:10:49 +0000 (14:10 -0700)]
Call get-magic once for implicit rv2gv in close(), etc.
This commit stops an implicit rv2gv from calling get-magic twice. As
a side-effect, it also squelches the duplicate warning emitted by
‘close undef’ (bug #97482).
is_gv_magical_sv is modified not to call get-magic on the sv passed to
it. It is not in the public API, and the only two callers (rv2gv and
softrefxv) have already called get-magic before calling it.