| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| |
|
|
|
|
|
|
|
|
|
|
| |
[DELTA]
VERSION 3.18 (2011-08-31)
NEW CODE(s)
No longer use CIA data
The CIA world added non-standard values, so I no longer use it as a source of data. Based on a report by Michiel Beijen.
|
| |
|
|
|
|
|
|
|
|
|
|
| |
Originally, GVs always had something in the SV slot. So, when the
code for $] and $^V started replacing it with another SV, it had to
free the existing SV.
Then commit c69033f2 came along and added the PERL_DONT_CREATE_GVSV
directive. It necessarily changed a bunch of GvSV()s to GvSVn()s in
gv_fetchpvn_flags. But it changed these two, even though they didn’t
need it. So, when PERL_DONT_CREATE_GVSV is true (the default), we
just create and throw away a scalar needlessly.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This resolves perl bug #97978.
Many built-in variables, like $], are actually created on the fly
when first accessed. Perl likes to pretend that these variables have
always existed, so it autovivifies the *] glob even in rvalue context
(e.g., defined *{"]"}, close "]").
The list of variables that were autovivified was maintained separ-
ately (in is_gv_magical_sv) from the code that actually creates
them (gv_fetchpvn_flags). ‘Maintained’ is not actually precise: it
*wasn’t* being maintained, and there were new variables that never
got added to is_gv_magical_sv and one deleted variable that was
never removed.
There are only two pieces of code that call is_gv_magical_sv, both in
pp.c: S_rv2gv (called by *{} and also the implicit *{} that functions
like close() provide) and Perl_softrefxv (called by ${}, @{}, %{}).
In both cases, the glob is immediately autovivified if
is_gv_magical_sv returns true.
So this commit eliminates the extra maintenance burden by extirpat-
ing is_gv_magical_sv altogether, and replacing it with a new flag to
gv_fetchpvn_flags, GvADDMG, which will autovivify a glob *if* it’s a
magical one.
It does make defined(*{"frobbly"}) slightly slower, in that it creates
a temporary glob and then frees it when it sees nothing magical has
been done with it. But this case is rare enough it should not matter.
At least I got rid of the bugginess.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Porting/findvars was added in 1998 in 2bd2b9e04a68ec86. It searches @ARGV for
its wordlist of then-current interpreter variable names.
Porting/fixvars was added as fixvars in 1998 in a15299417de39f35. It captures
the output of make (defaulting to make miniperl), parses it for errors
matching /undeclared/, and then attempts to edit the relevant line of the
reported file to prefix the name with PL_
Porting/fixCORE is a modified copy of Porting/fixvars, added in 1998 in
a8693bd382efcc6d. It's intended to load modules, catching errors of the form
/Ambiguous call resolved as CORE::/, and editing the relevant lines to prefix
CORE:: to the function in question. It appears only ever to have been used to
fix warnings in Math::Complex.
All 3 have only had trivial style and spelling edits since addition, and
have been unused for over 12 years.
|
| | |
|
| |
|
|
|
| |
The fact that I typed a comma for 2 probably bewrays what keyboard
layout I’m using. :-)
|
| | |
|
| |
|
|
|
|
| |
This commit allows &CORE::write to be called through references and
via ampersand syntax. No change to pp_enterwrite was necessary, as it
can already handle nulls.
|
| |
|
|
|
|
|
|
|
|
|
| |
This commit allows &CORE::unpack to be called through references and
via ampersand syntax.
It moves the $_-handling code in pp_coreargs inside the parameter
loop, so it can apply to the second parameter, not just the first.
Consequently, a mkdir test has been added that ensures implicit $_
is not used for mkdir’s second argument; i.e., that the $_-handling
code’s if() condition is correct.
|
| |
|
|
|
|
|
|
| |
This commit allows &CORE::umask to be called through references and
via ampersand syntax. pp_umask 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 umask how many arguments it’s
actually getting. See commit 0163043a for details.
|
| |
|
|
|
|
|
| |
This commit allows the tie, tied and untie subroutines in the CORE
namespace to be called through references and via &ersand() syntax.
pp_coreargs is modified to handle the functions with \[$@%*] in their
prototypes (which happen to be just the tie functions).
|
| |
|
|
|
|
|
|
| |
This commit allows &CORE::tell to be called through references and
via ampersand syntax. pp_tell 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_tell how many arguments it’s
actually getting. See commit 0163043a for details.
|
| |
|
|
|
|
|
|
| |
This commit allows &CORE::setpgrp to be called through references and
via ampersand syntax. pp_setpgrp 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 setpgrp how many arguments it’s
actually getting. See commit 0163043a for details.
|
| |
|
|
|
| |
Prior to this commit, setpgrp(27) was equivalent to (27, setpgrp),
because it ignored the argument on the stack when there was only one.
|
| | |
|
| |
|
|
|
| |
... file when -except is used; write the '\0' escape sequence
properly instead.
|
| | |
|
| |
|
|
|
|
|
|
| |
8ca8a454f60a417f optimised the implementation of s///r by avoiding an
unconditional copy of the original value. However, it introduced a behaviour
regression where if original value happened to be one of a few particular
types, it could be modified by being forced to a string using SvPV_force().
The substitution was (correctly) performed on a copy of this string.
|
| | |
|
| | |
|
| | |
|
| |
|
|
|
| |
This way we get to make use of PAUSE's permission system instead of allowing
everyone to to upload new perlfaq versions.
|
| |
|
|
|
|
|
|
| |
This commit allows &CORE::sysopen to be called through references and
via ampersand syntax. pp_sysopen 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 sysopen how many arguments it’s
actually getting. See commit 0163043a for details.
|
| |
|
|
|
|
| |
and not just lock, vec and substr. Using a regular leavesub op causes
the return values to be copied. There is no need for that, so this
commit changes them all to use leavesublv.
|
| | |
|
| | |
|
| |
|
|
|
|
|
|
|
|
| |
Turns out comparing a signed byte to values above 0x7f doesn't make
sense. *All* signed byte integers are less than or equal to 0x9f,
so the other two branches of the if could never be taken.
This code probably needs more review and testing, but we might as
well make it do what it intends to do before reviewing those
intentions and factoring out some of the copy-and-paste verbosity.
|
| |
|
|
|
|
|
|
| |
With XS(name) defaulting to exporting symbols again since the
previous commit, ExtUtils::ParseXS will now instead use explicit
XS_EXTERNAL/XS_INTERNAL in its place. This allows backporting
of the linkage changes to perls as old as 5.10.0 (and possibly
further).
|
| |
|
|
|
|
|
| |
Instead, as Zefram recommended, ExtUtils::ParseXS will be patched
to not export XSUB symbols by default that are generated through
the module itself. As Zefram said, this has the advantage of
allowing older perls to benefit from the non-exporting of symbols.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This commit makes &CORE::substr callable through references and via
&ersand syntax.
It’s a bit awkward, as we need a substr op that is flagged as hav-
ing 4 arguments *and* possibly returning an lvalue. The code in
op_lvalue_flags wasn’t really set up for that, so I needed to flag
the op with OPpMAYBE_LVSUB in coresub_op before it gets passed to
op_lvalue_flags. It turned out that only that was necessary, as
op_lvalue_flags does an op_private == 4 check (rather than (op_private
& 7) == 4 or some such) when checking for the 4-arg case and croak-
ing. When the op arrives in op_lvalue_flags, it’s already flagged
OPpMAYBE_LVSUB|4 which != 4.
pp_substr is also modified to check for nulls and, if necessary,
adjust its count of how many arguments were actually passed.)
|
| |
|
|
|
|
|
|
| |
This commit allows &CORE::srand to be called through references and
via ampersand syntax. pp_srand 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 srand how many arguments it’s
actually getting. See commit 0163043a for details.
|
| |
|
|
|
|
|
|
|
|
|
| |
This commit allows &CORE::sleep to be called through references and
via ampersand syntax. pp_sleep 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 sleep how many arguments it’s
actually getting. See commit 0163043a for details.
Unfortunately, sleep with no args is nearly impossible to test porta-
bly. But I have checked that it works.
|
| |
|
|
|
|
|
|
|
|
| |
Clearly this code is not well-tested. Especially the parts that
the compiler couldn't see because it was hidden by unclosed
comment blocks. Ouch.
I'm not going to name names or look too deep in the blame log, but
suffice it to say that this is why patches that add hundreds of
lines of new code without tests should not be accepted.
|
| |
|
|
|
|
|
|
|
| |
A couple of empty parameter lists and a couple of unsigned -1
error statuses. Brought to you by
-Duser_c_flags=/WARN=(ENABLE=LEVEL4,INFORMATIONAL=ALL)
which is only the third highest warning level.
|
| |
|
|
|
| |
A couple are only needed for the homegrown utime() replacement on
older systems. rmsdef.h is apparently not needed at all.
|
| |
|
|
|
|
|
|
| |
This commit makes &CORE::send and &CORE::syswrite callable through references and & syntax.
All this commit has to do is remove them from the exception list in
gv.c, as previous commits happen to have made them work. (I didn’t
realise originally that these use pushmark.)
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This commit allows CORE::select to be called through references and
via &ersand syntax.
This is a tricky case, as the select keyword represents two distinct
operators. ck_select replaces the OP_SELECT with an OP_SSELECT if
there is more that one argument.
So what we do here is create an if(@_>1)/else block with the usual
op-with-coreargs-child inside each branch.
The op tree looks like this:
$ ./perl -Ilib -mO=Concise,CORE::select -e 'BEGIN{\&CORE::select}
'
CORE::select:
8 <1> leavesub[1 ref] K/REFC,1 ->(end)
- <1> null K/1 ->8
5 <|> cond_expr(other->6) K/1 ->9
4 <2> gt sK/2 ->5
2 <1> rv2av[t4] sK/1 ->3
1 <#> gv[*_] s ->2
3 <$> const[IV 1] s ->4
7 <@> sselect[t2] K ->8
- <0> ex-pushmark s ->6
6 <$> coreargs(IV 218) ->7
a <@> select[t1] sK/1 ->8
- <0> ex-pushmark s ->9
9 <$> coreargs(IV 219) s/DREF1 ->a
-e syntax OK
There was no need to modify pp_select to handle a null when there is
no argument, as it can already handle it.
|
| | |
|
| | |
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
| |
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.
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
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.
|
| |
|
|
|
| |
This now reflects Tom Christiansen's and my current thinking about Greek
Final Sigma
|
| |
|
|
|
|
| |
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.
|
| | |
|
| |
|
|
|
|
|
|
| |
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.
|
| |
|
|
|
|
|
|
|
|
| |
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
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
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.
|
| |
|
|
|
| |
pp_coreargs will use this to distinguish between the \$ and \[$@%*]
prototypes.
|
| | |
|
