SYNOPSIS
use Data::Peek;
print DDumper \%hash; # Same syntax as Data::Dumper
DTidy { ref => $ref };
print DPeek \$var;
my ($pv, $iv, $nv, $rv, $magic) = DDual ($var [, 1]);
print DPeek for DDual ($!, 1);
print DDisplay ("ab\nc\x{20ac}\rdef\n");
print DHexDump ("ab\nc\x{20ac}\rdef\n");
my $dump = DDump $var;
my %hash = DDump \@list;
DDump \%hash;
my %hash = DDump (\%hash, 5); # dig 5 levels deep
my $dump;
open my $fh, ">", \$dump;
DDump_IO ($fh, \%hash, 6);
close $fh;
print $dump;
# Imports
use Data::Peek qw( :tidy VNR DGrow triplevar );
my $x = ""; DGrow ($x, 10000);
my $tv = triplevar ("\N{GREEK SMALL LETTER PI}", 3, "3.1415");
DDsort ("R");
DDumper [ $x ]; # use of :tidy make DDumper behave as DTidy
DESCRIPTION
Data::Peek started off as "DDumper" being a wrapper module over Data::Dumper, but grew out to be a set of low-level data introspection utilities that no other module provided yet, using the lowest level of the perl internals API as possible.DDumper ($var, ...)
Not liking the default output of Data::Dumper, and always feeling the need to set "$Data::Dumper::Sortkeys = 1;", and not liking any of the default layouts, this function is just a wrapper around Data::Dumper::Dumper with everything set as I like it.
$Data::Dumper::Sortkeys = 1;
$Data::Dumper::Indent = 1;
If "Data::Peek" is "use"d with import argument ":tidy", the result is formatted according to Perl::Tidy, see DTidy below, otherwise the result is further beautified to meet my needs:
* quotation of hash keys has been removed (with the disadvantage
that the output might not be parseable again).
* arrows for hashes are aligned at 16 (longer keys don't align)
* closing braces and brackets are now correctly aligned
In void context, "DDumper ()" warn ()'s.
Example
$ perl -MDP \
-e'DDumper { ape => 1, foo => "egg", bar => [ 2, "baz", undef ]};'
{ ape => 1,
bar => [
2,
'baz',
undef
],
foo => 'egg'
};
DTidy ($var, ...)
"DTidy" is an alternative to "DDumper", where the output of "DDumper" is formatted using "Perl::Tidy" (if available) according to your ".perltidyrc" instead of the default behavior, maybe somewhat like (YMMV):
$ perl -MDP=:tidy \
-we'DDumper { ape => 1, foo => "egg", bar => [ 2, "baz", undef ]};'
{ 'ape' => 1,
'bar' => [2, 'baz', undef],
'foo' => 'egg'
}
If "Data::Peek" is "use"d with import argument ":tidy", this is the default output method for "DDumper".
If Perl::Tidy is not available, "DTidy" will fallback to "DDumper".
This idea was shamelessly copied from John McNamara's Data::Dumper::Perltidy.
DDsort ( 0 | 1 | R | V | VR | VN | VNR )
Set the hash sort algorithm for DDumper. The default is to sort by key value.
0 - Do not sort 1 - Sort by key R - Reverse sort by key V - Sort by value VR - Reverse sort by value VN - Sort by value numerical VNR - Reverse sort by value numerical
These can also be passed to import:
$ perl -MDP=VNR \
-we'DDumper { foo => 1, bar => 2, zap => 3, gum => 13 }'
{ gum => 13,
zap => 3,
bar => 2,
foo => 1
};
$ perl -MDP=V \
-we'DDumper { foo => 1, bar => 2, zap => 3, gum => 13 }'
{ foo => 1,
gum => 13,
bar => 2,
zap => 3
};
DPeek
DPeek ($var)
Playing with "sv_dump ()", I found "Perl_sv_peek ()", and it might be very useful for simple checks. If $var is omitted, uses $_.Example
print DPeek "abc\x{0a}de\x{20ac}fg";
PV("abc\nde\342\202\254fg"\0) [UTF8 "abc\nde\x{20ac}fg"]
In void context, "DPeek ()" prints to "STDERR" plus a newline.
DDisplay
DDisplay ($var)
Show the PV content of a scalar the way perl debugging would have done. UTF-8 detection is on, so this is effectively the same as returning the first part the "DPeek ()" returns for non-UTF8 PV's or the second part for UTF-8 PV's. "DDisplay ()" returns the empty string for scalars that no have a valid PV.Example
print DDisplay "abc\x{0a}de\x{20ac}fg";
"abc\nde\x{20ac}fg"
In void context, "DDisplay ()" uses "warn" to display the result.
DHexDump
DHexDump ($var)
DHexDump ($var, $length)
Show the (stringified) content of a scalar as a hex-dump. If $var is omitted, $_ is dumped. Returns "undef" or an empty list if $var (or $_) is undefined. If $length is given and is lower than the length of the stringified $var, only <$length> bytes are dumped.In void context, the dump is done to STDERR. In scalar context, the complete dump is returned as a single string. In list context, the dump is returned as lines.
Example
print DHexDump "abc\x{0a}de\x{20ac}fg";
0000 61 62 63 0a 64 65 e2 82 ac 66 67 abc.de...fg
my ($pv, $iv, $nv, $rv, $hm) = DDual ($var [, $getmagic])
DDual will return the basic elements in a variable, guaranteeing that no conversion takes place. This is very useful for dual-var variables, or when checking is a variable has defined entries for a certain type of scalar. For each String (PV), Integer (IV), Double (NV), and Reference (RV), the current value of $var is returned or undef if it is not set (yet). The 5th element is an indicator if $var has magic, which is not invoked in the returned values, unless explicitly asked for with a true optional second argument.Example
print DPeek for DDual ($!, 1);
In void context, DDual does the equivalent of
{ my @d = DDual ($!, 1);
print STDERR
DPeek ($!), "\n",
" PV: ", DPeek ($d[0]), "\n",
" IV: ", DPeek ($d[1]), "\n",
" NV: ", DPeek ($d[2]), "\n",
" RV: ", DPeek ($d[3]), "\n";
}
my $len = DGrow ($pv, $size)
Fastest way to preallocate space for a PV scalar. Returns the allocated length. If $size is smaller than the already allocated space, it will not shrink.
cmpthese (-2, {
pack => q{my $x = ""; $x = pack "x20000"; $x = "";},
op_x => q{my $x = ""; $x = "x" x 20000; $x = "";},
grow => q{my $x = ""; DGrow ($x, 20000); $x = "";},
});
Rate op_x pack grow 5.8.9 5.10.1 5.12.4 5.14.2
op_x 62127/s -- -59% -96% 118606/s 119730/s 352255/s 362605/s
pack 152046/s 145% -- -91% 380075/s 355666/s 347247/s 387349/s
grow 1622943/s 2512% 967% -- 2818380/s 2918783/s 2672340/s 2886787/s
my $tp = triplevar ($pv, $iv, $nv)
When making "DDual ()" I wondered if it were possible to create triple-val scalar variables. Scalar::Util already gives us "dualvar ()", that creates you a scalar with different numeric and string values that return different values in different context. Not that "triplevar ()" would be very useful, compared to "dualvar ()", but at least this shows that it is possible."triplevar ()" is not exported by default.
Example:
print DPeek for DDual
Data::Peek::triplevar ("\N{GREEK SMALL LETTER PI}", 3, 3.1415);
PV("\317\200"\0) [UTF8 "\x{3c0}"]
IV(3)
NV(3.1415)
SV_UNDEF
IV(0)
DDump ($var [, $dig_level])
A very useful module when debugging is "Devel::Peek", but is has one big disadvantage: it only prints to STDERR, which is not very handy when your code wants to inspect variables at a low level.Perl itself has "sv_dump ()", which does something similar, but still prints to STDERR, and only one level deep.
"DDump ()" is an attempt to make the innards available to the script level with a reasonable level of compatibility. "DDump ()" is context sensitive.
In void context, it behaves exactly like "Perl_sv_dump ()".
In scalar context, it returns what "Perl_sv_dump ()" would have printed.
In list context, it returns a hash of the variable's properties. In this mode you can pass an optional second argument that determines the depth of digging.
Example
print scalar DDump "abc\x{0a}de\x{20ac}fg"
SV = PV(0x723250) at 0x8432b0
REFCNT = 1
FLAGS = (PADBUSY,PADMY,POK,pPOK,UTF8)
PV = 0x731ac0 "abc\nde\342\202\254fg"\0 [UTF8 "abc\nde\x{20ac}fg"]
CUR = 11
LEN = 16
my %h = DDump "abc\x{0a}de\x{20ac}fg";
print DDumper \%h;
{ CUR => '11',
FLAGS => {
PADBUSY => 1,
PADMY => 1,
POK => 1,
UTF8 => 1,
pPOK => 1
},
LEN => '16',
PV => '0x731ac0 "abc\\nde\\342\\202\\254fg"\\0 [UTF8 "abc\\nde\\x{20ac}fg"]',
REFCNT => '1',
sv => 'PV(0x723250) at 0x8432c0'
};
my %h = DDump {
ape => 1,
foo => "egg",
bar => [ 2, "baz", undef ],
}, 1;
print DDumper \%h;
{ FLAGS => {
PADBUSY => 1,
PADMY => 1,
ROK => 1
},
REFCNT => '1',
RV => {
PVIV("ape") => {
FLAGS => {
IOK => 1,
PADBUSY => 1,
PADMY => 1,
pIOK => 1
},
IV => '1',
REFCNT => '1',
sv => 'IV(0x747020) at 0x843a10'
},
PVIV("bar") => {
CUR => '0',
FLAGS => {
PADBUSY => 1,
PADMY => 1,
ROK => 1
},
IV => '1',
LEN => '0',
PV => '0x720210 ""',
REFCNT => '1',
RV => '0x720210',
sv => 'PVIV(0x7223e0) at 0x843a10'
},
PVIV("foo") => {
CUR => '3',
FLAGS => {
PADBUSY => 1,
PADMY => 1,
POK => 1,
pPOK => 1
},
IV => '1',
LEN => '8',
PV => '0x7496c0 "egg"\\0',
REFCNT => '1',
sv => 'PVIV(0x7223e0) at 0x843a10'
}
},
sv => 'RV(0x79d058) at 0x843310'
};
DDump_IO ($io, $var [, $dig_level])
A wrapper function around perl's internal "Perl_do_sv_dump ()", which makes "Devel::Peek" completely superfluous.Example
my $dump;
open my $eh, ">", \$dump;
DDump_IO ($eh, { 3 => 4, ape => [5..8]}, 6);
close $eh;
print $dump;
SV = RV(0x79d9e0) at 0x843f00
REFCNT = 1
FLAGS = (TEMP,ROK)
RV = 0x741090
SV = PVHV(0x79c948) at 0x741090
REFCNT = 1
FLAGS = (SHAREKEYS)
IV = 2
NV = 0
ARRAY = 0x748ff0 (0:7, 2:1)
hash quality = 62.5%
KEYS = 2
FILL = 1
MAX = 7
RITER = -1
EITER = 0x0
Elt "ape" HASH = 0x97623e03
SV = RV(0x79d9d8) at 0x8440e0
REFCNT = 1
FLAGS = (ROK)
RV = 0x741470
SV = PVAV(0x7264b0) at 0x741470
REFCNT = 2
FLAGS = ()
IV = 0
NV = 0
ARRAY = 0x822f70
FILL = 3
MAX = 3
ARYLEN = 0x0
FLAGS = (REAL)
Elt No. 0
SV = IV(0x7467c8) at 0x7c1aa0
REFCNT = 1
FLAGS = (IOK,pIOK)
IV = 5
Elt No. 1
SV = IV(0x7467b0) at 0x8440f0
REFCNT = 1
FLAGS = (IOK,pIOK)
IV = 6
Elt No. 2
SV = IV(0x746810) at 0x75be00
REFCNT = 1
FLAGS = (IOK,pIOK)
IV = 7
Elt No. 3
SV = IV(0x746d38) at 0x7799d0
REFCNT = 1
FLAGS = (IOK,pIOK)
IV = 8
Elt "3" HASH = 0xa400c7f3
SV = IV(0x746fd0) at 0x7200e0
REFCNT = 1
FLAGS = (IOK,pIOK)
IV = 4
INTERNALS
"DDump ()" uses an XS wrapper around "Perl_sv_dump ()" where the STDERR is temporarily caught to a pipe. The internal XS helper functions are not meant for user spaceDDump_XS (SV *sv)
Base interface to internals for "DDump ()".BUGS
Windows and AIX might be using a build where not all symbols that were supposed to be exported in the public API are not. Perl_pv_peek () is one of them.Not all types of references are supported.
No idea how far back this goes in perl support, but Devel::PPPort has proven to be a big help.
AUTHOR
H.Merijn Brand <[email protected]>COPYRIGHT AND LICENSE
Copyright (C) 2008-2016 H.Merijn BrandThis library is free software; you can redistribute it and/or modify it under the same terms as Perl itself.