| line |
stmt |
bran |
cond |
sub |
pod |
time |
code |
|
1
|
|
|
|
|
|
|
|
|
2
|
|
|
|
|
|
|
|
|
3
|
|
|
|
|
|
|
|
|
4
|
|
|
|
|
|
|
|
|
5
|
|
|
|
|
|
|
|
|
6
|
|
|
|
|
|
|
|
|
7
|
|
|
|
|
|
|
|
|
8
|
|
|
|
|
|
|
|
|
9
|
|
|
|
|
|
|
|
|
10
|
|
|
|
|
|
|
|
|
11
|
|
|
|
|
|
|
|
|
12
|
|
|
|
|
|
|
|
|
13
|
|
|
|
|
|
|
|
|
14
|
|
|
|
|
|
|
|
|
15
|
|
|
|
|
|
|
|
|
16
|
|
|
|
|
|
|
|
|
17
|
|
|
|
|
|
|
|
|
18
|
|
|
|
|
|
|
|
|
19
|
|
|
|
|
|
|
|
|
20
|
|
|
|
|
|
|
|
|
21
|
|
|
|
|
|
|
|
|
22
|
|
|
|
|
|
|
|
|
23
|
|
|
|
|
|
|
|
|
24
|
|
|
|
|
|
|
|
|
25
|
|
|
|
|
|
|
package Convert::Binary::C; |
|
26
|
|
|
|
|
|
|
|
|
27
|
61
|
|
|
61
|
|
954
|
use strict; |
|
|
61
|
|
|
|
|
1673
|
|
|
|
61
|
|
|
|
|
1836
|
|
|
28
|
61
|
|
|
61
|
|
1064
|
use DynaLoader; |
|
|
61
|
|
|
|
|
2793
|
|
|
|
61
|
|
|
|
|
1716
|
|
|
29
|
61
|
|
|
61
|
|
1353
|
use Carp; |
|
|
61
|
|
|
|
|
546
|
|
|
|
61
|
|
|
|
|
1109
|
|
|
30
|
61
|
|
|
61
|
|
919
|
use vars qw( @ISA $VERSION $XS_VERSION $AUTOLOAD ); |
|
|
61
|
|
|
|
|
614
|
|
|
|
61
|
|
|
|
|
3554
|
|
|
31
|
|
|
|
|
|
|
|
|
32
|
|
|
|
|
|
|
@ISA = qw(DynaLoader); |
|
33
|
|
|
|
|
|
|
|
|
34
|
|
|
|
|
|
|
$VERSION = do { my @r = '$Snapshot: /Convert-Binary-C/0.67 $' =~ /(\d+\.\d+(?:_\d+)?)/; @r ? $r[0] : '9.99' }; |
|
35
|
|
|
|
|
|
|
$XS_VERSION = $VERSION; |
|
36
|
|
|
|
|
|
|
$VERSION = eval $VERSION; |
|
37
|
|
|
|
|
|
|
|
|
38
|
|
|
|
|
|
|
bootstrap Convert::Binary::C $XS_VERSION; |
|
39
|
|
|
|
|
|
|
|
|
40
|
|
|
|
|
|
|
|
|
41
|
|
|
|
|
|
|
|
|
42
|
|
|
|
|
|
|
|
|
43
|
|
|
|
|
|
|
sub AUTOLOAD |
|
44
|
|
|
|
|
|
|
{ |
|
45
|
6200
|
|
|
6200
|
|
11589590
|
my $self = shift; |
|
46
|
6200
|
|
|
|
|
74657
|
my $opt = $AUTOLOAD; |
|
47
|
6200
|
50
|
|
|
|
303279
|
ref $self or croak "$self is not an object"; |
|
48
|
6200
|
|
|
|
|
159911
|
$opt =~ s/.*://; |
|
49
|
6200
|
100
|
|
|
|
178656
|
$opt =~ /^[A-Z]/ or croak "Invalid method $opt called"; |
|
50
|
6199
|
100
|
|
|
|
135261
|
@_ <= 1 or croak "$opt cannot take more than one argument"; |
|
51
|
6198
|
100
|
100
|
|
|
80659
|
unless (@_ or defined wantarray) { |
|
52
|
21
|
|
|
|
|
279
|
carp "Useless use of $opt in void context"; |
|
53
|
21
|
|
|
|
|
580
|
return; |
|
54
|
|
|
|
|
|
|
} |
|
55
|
6177
|
|
|
|
|
66729
|
my @warn; |
|
56
|
|
|
|
|
|
|
{ |
|
57
|
6177
|
|
|
4
|
|
62913
|
local $SIG{__WARN__} = sub { push @warn, $_[0] }; |
|
|
6177
|
|
|
|
|
213388
|
|
|
|
4
|
|
|
|
|
51
|
|
|
58
|
6177
|
|
|
|
|
129738
|
$opt = eval { $self->configure( $opt, @_ ) }; |
|
|
6177
|
|
|
|
|
157486
|
|
|
59
|
|
|
|
|
|
|
} |
|
60
|
6177
|
|
|
|
|
242186
|
for my $w (@warn) { |
|
61
|
4
|
|
|
|
|
83
|
$w =~ s/\s+at.*?C\.pm.*//s; |
|
62
|
4
|
|
|
|
|
54
|
carp $w; |
|
63
|
|
|
|
|
|
|
} |
|
64
|
6177
|
100
|
|
|
|
75689
|
if ($@) { |
|
65
|
198
|
|
|
|
|
3458
|
$@ =~ s/\s+at.*?C\.pm.*//s; |
|
66
|
198
|
|
|
|
|
2448
|
croak $@; |
|
67
|
|
|
|
|
|
|
} |
|
68
|
5979
|
|
|
|
|
87570
|
$opt; |
|
69
|
|
|
|
|
|
|
} |
|
70
|
|
|
|
|
|
|
|
|
71
|
|
|
|
|
|
|
1; |
|
72
|
|
|
|
|
|
|
|
|
73
|
|
|
|
|
|
|
__END__ |
|
74
|
|
|
|
|
|
|
|
|
75
|
|
|
|
|
|
|
=head1 NAME |
|
76
|
|
|
|
|
|
|
|
|
77
|
|
|
|
|
|
|
Convert::Binary::C - Binary Data Conversion using C Types |
|
78
|
|
|
|
|
|
|
|
|
79
|
|
|
|
|
|
|
=head1 SYNOPSIS |
|
80
|
|
|
|
|
|
|
|
|
81
|
|
|
|
|
|
|
=head2 Simple |
|
82
|
|
|
|
|
|
|
|
|
83
|
|
|
|
|
|
|
use Convert::Binary::C; |
|
84
|
|
|
|
|
|
|
|
|
85
|
|
|
|
|
|
|
#--------------------------------------------- |
|
86
|
|
|
|
|
|
|
# Create a new object and parse embedded code |
|
87
|
|
|
|
|
|
|
#--------------------------------------------- |
|
88
|
|
|
|
|
|
|
my $c = Convert::Binary::C->new->parse(<<ENDC); |
|
89
|
|
|
|
|
|
|
|
|
90
|
|
|
|
|
|
|
enum Month { JAN, FEB, MAR, APR, MAY, JUN, |
|
91
|
|
|
|
|
|
|
JUL, AUG, SEP, OCT, NOV, DEC }; |
|
92
|
|
|
|
|
|
|
|
|
93
|
|
|
|
|
|
|
struct Date { |
|
94
|
|
|
|
|
|
|
int year; |
|
95
|
|
|
|
|
|
|
enum Month month; |
|
96
|
|
|
|
|
|
|
int day; |
|
97
|
|
|
|
|
|
|
}; |
|
98
|
|
|
|
|
|
|
|
|
99
|
|
|
|
|
|
|
ENDC |
|
100
|
|
|
|
|
|
|
|
|
101
|
|
|
|
|
|
|
#----------------------------------------------- |
|
102
|
|
|
|
|
|
|
# Pack Perl data structure into a binary string |
|
103
|
|
|
|
|
|
|
#----------------------------------------------- |
|
104
|
|
|
|
|
|
|
my $date = { year => 2002, month => 'DEC', day => 24 }; |
|
105
|
|
|
|
|
|
|
|
|
106
|
|
|
|
|
|
|
my $packed = $c->pack('Date', $date); |
|
107
|
|
|
|
|
|
|
|
|
108
|
|
|
|
|
|
|
=head2 Advanced |
|
109
|
|
|
|
|
|
|
|
|
110
|
|
|
|
|
|
|
use Convert::Binary::C; |
|
111
|
|
|
|
|
|
|
use Data::Dumper; |
|
112
|
|
|
|
|
|
|
|
|
113
|
|
|
|
|
|
|
#--------------------- |
|
114
|
|
|
|
|
|
|
# Create a new object |
|
115
|
|
|
|
|
|
|
#--------------------- |
|
116
|
|
|
|
|
|
|
my $c = new Convert::Binary::C ByteOrder => 'BigEndian'; |
|
117
|
|
|
|
|
|
|
|
|
118
|
|
|
|
|
|
|
#--------------------------------------------------- |
|
119
|
|
|
|
|
|
|
# Add include paths and global preprocessor defines |
|
120
|
|
|
|
|
|
|
#--------------------------------------------------- |
|
121
|
|
|
|
|
|
|
$c->Include('/usr/lib/gcc-lib/i686-pc-linux-gnu/3.3.6/include', |
|
122
|
|
|
|
|
|
|
'/usr/include') |
|
123
|
|
|
|
|
|
|
->Define(qw( __USE_POSIX __USE_ISOC99=1 )); |
|
124
|
|
|
|
|
|
|
|
|
125
|
|
|
|
|
|
|
#---------------------------------- |
|
126
|
|
|
|
|
|
|
# Parse the 'time.h' header file |
|
127
|
|
|
|
|
|
|
#---------------------------------- |
|
128
|
|
|
|
|
|
|
$c->parse_file('time.h'); |
|
129
|
|
|
|
|
|
|
|
|
130
|
|
|
|
|
|
|
#--------------------------------------- |
|
131
|
|
|
|
|
|
|
# See which files the object depends on |
|
132
|
|
|
|
|
|
|
#--------------------------------------- |
|
133
|
|
|
|
|
|
|
print Dumper([$c->dependencies]); |
|
134
|
|
|
|
|
|
|
|
|
135
|
|
|
|
|
|
|
#----------------------------------------------------------- |
|
136
|
|
|
|
|
|
|
# See if struct timespec is defined and dump its definition |
|
137
|
|
|
|
|
|
|
#----------------------------------------------------------- |
|
138
|
|
|
|
|
|
|
if ($c->def('struct timespec')) { |
|
139
|
|
|
|
|
|
|
print Dumper($c->struct('timespec')); |
|
140
|
|
|
|
|
|
|
} |
|
141
|
|
|
|
|
|
|
|
|
142
|
|
|
|
|
|
|
#------------------------------- |
|
143
|
|
|
|
|
|
|
# Create some binary dummy data |
|
144
|
|
|
|
|
|
|
#------------------------------- |
|
145
|
|
|
|
|
|
|
my $data = "binary_test_string"; |
|
146
|
|
|
|
|
|
|
|
|
147
|
|
|
|
|
|
|
#-------------------------------------------------------- |
|
148
|
|
|
|
|
|
|
# Unpack $data according to 'struct timespec' definition |
|
149
|
|
|
|
|
|
|
#-------------------------------------------------------- |
|
150
|
|
|
|
|
|
|
if (length($data) >= $c->sizeof('timespec')) { |
|
151
|
|
|
|
|
|
|
my $perl = $c->unpack('timespec', $data); |
|
152
|
|
|
|
|
|
|
print Dumper($perl); |
|
153
|
|
|
|
|
|
|
} |
|
154
|
|
|
|
|
|
|
|
|
155
|
|
|
|
|
|
|
#-------------------------------------------------------- |
|
156
|
|
|
|
|
|
|
# See which member lies at offset 5 of 'struct timespec' |
|
157
|
|
|
|
|
|
|
#-------------------------------------------------------- |
|
158
|
|
|
|
|
|
|
my $member = $c->member('timespec', 5); |
|
159
|
|
|
|
|
|
|
print "member('timespec', 5) = '$member'\n"; |
|
160
|
|
|
|
|
|
|
|
|
161
|
|
|
|
|
|
|
=head1 DESCRIPTION |
|
162
|
|
|
|
|
|
|
|
|
163
|
|
|
|
|
|
|
Convert::Binary::C is a preprocessor and parser for C type |
|
164
|
|
|
|
|
|
|
definitions. It is highly configurable and supports |
|
165
|
|
|
|
|
|
|
arbitrarily complex data structures. Its object-oriented |
|
166
|
|
|
|
|
|
|
interface has L<C<pack>|/"pack"> and L<C<unpack>|/"unpack"> methods |
|
167
|
|
|
|
|
|
|
that act as replacements for |
|
168
|
|
|
|
|
|
|
Perl's L<C<pack>|perlfunc/"pack"> and L<C<unpack>|perlfunc/"unpack"> and |
|
169
|
|
|
|
|
|
|
allow to use C types instead of a string representation |
|
170
|
|
|
|
|
|
|
of the data structure for conversion of binary data from and |
|
171
|
|
|
|
|
|
|
to Perl's complex data structures. |
|
172
|
|
|
|
|
|
|
|
|
173
|
|
|
|
|
|
|
Actually, what Convert::Binary::C does is not very different |
|
174
|
|
|
|
|
|
|
from what a C compiler does, just that it doesn't compile the |
|
175
|
|
|
|
|
|
|
source code into an object file or executable, but only parses |
|
176
|
|
|
|
|
|
|
the code and allows Perl to use the enumerations, structs, unions |
|
177
|
|
|
|
|
|
|
and typedefs that have been defined within your C source for binary |
|
178
|
|
|
|
|
|
|
data conversion, similar to |
|
179
|
|
|
|
|
|
|
Perl's L<C<pack>|perlfunc/"pack"> and L<C<unpack>|perlfunc/"unpack">. |
|
180
|
|
|
|
|
|
|
|
|
181
|
|
|
|
|
|
|
Beyond that, the module offers a lot of convenience methods |
|
182
|
|
|
|
|
|
|
to retrieve information about the C types that have been parsed. |
|
183
|
|
|
|
|
|
|
|
|
184
|
|
|
|
|
|
|
=head2 Background and History |
|
185
|
|
|
|
|
|
|
|
|
186
|
|
|
|
|
|
|
In late 2000 I wrote a real-time debugging interface for an |
|
187
|
|
|
|
|
|
|
embedded medical device that allowed me to send out data from |
|
188
|
|
|
|
|
|
|
that device over its integrated Ethernet adapter. |
|
189
|
|
|
|
|
|
|
The interface was C<printf()>-like, so you could easily send |
|
190
|
|
|
|
|
|
|
out strings or numbers. But you could also send out what I |
|
191
|
|
|
|
|
|
|
called I<arbitrary data>, which was intended for arbitrary |
|
192
|
|
|
|
|
|
|
blocks of the device's memory. |
|
193
|
|
|
|
|
|
|
|
|
194
|
|
|
|
|
|
|
Another part of this real-time debugger was a Perl application |
|
195
|
|
|
|
|
|
|
running on my workstation that gathered all the messages that |
|
196
|
|
|
|
|
|
|
were sent out from the embedded device. It printed all the |
|
197
|
|
|
|
|
|
|
strings and numbers, and hex-dumped the arbitrary data. |
|
198
|
|
|
|
|
|
|
However, manually parsing a couple of 300 byte hex-dumps of a |
|
199
|
|
|
|
|
|
|
complex C structure is not only frustrating, but also error-prone |
|
200
|
|
|
|
|
|
|
and time consuming. |
|
201
|
|
|
|
|
|
|
|
|
202
|
|
|
|
|
|
|
Using L<C<unpack>|perlfunc/"unpack"> to retrieve the contents |
|
203
|
|
|
|
|
|
|
of a C structure works fine for small structures and if you |
|
204
|
|
|
|
|
|
|
don't have to deal with struct member alignment. But otherwise, |
|
205
|
|
|
|
|
|
|
maintaining such code can be as awful as deciphering hex-dumps. |
|
206
|
|
|
|
|
|
|
|
|
207
|
|
|
|
|
|
|
As I didn't find anything to solve my problem on the CPAN, |
|
208
|
|
|
|
|
|
|
I wrote a little module that translated simple C structs |
|
209
|
|
|
|
|
|
|
into L<C<unpack>|perlfunc/"unpack"> strings. It worked, but |
|
210
|
|
|
|
|
|
|
it was slow. And since it couldn't deal with struct member |
|
211
|
|
|
|
|
|
|
alignment, I soon found myself adding padding bytes everywhere. |
|
212
|
|
|
|
|
|
|
So again, I had to maintain two sources, and changing one of |
|
213
|
|
|
|
|
|
|
them forced me to touch the other one. |
|
214
|
|
|
|
|
|
|
|
|
215
|
|
|
|
|
|
|
All in all, this little module seemed to make my task a bit |
|
216
|
|
|
|
|
|
|
easier, but it was far from being what I was thinking of: |
|
217
|
|
|
|
|
|
|
|
|
218
|
|
|
|
|
|
|
=over 2 |
|
219
|
|
|
|
|
|
|
|
|
220
|
|
|
|
|
|
|
=item * |
|
221
|
|
|
|
|
|
|
|
|
222
|
|
|
|
|
|
|
A module that could directly use the source I've been coding |
|
223
|
|
|
|
|
|
|
for the embedded device without any modifications. |
|
224
|
|
|
|
|
|
|
|
|
225
|
|
|
|
|
|
|
=item * |
|
226
|
|
|
|
|
|
|
|
|
227
|
|
|
|
|
|
|
A module that could be configured to match the properties |
|
228
|
|
|
|
|
|
|
of the different compilers and target platforms I was using. |
|
229
|
|
|
|
|
|
|
|
|
230
|
|
|
|
|
|
|
=item * |
|
231
|
|
|
|
|
|
|
|
|
232
|
|
|
|
|
|
|
A module that was fast enough to decode a great amount of |
|
233
|
|
|
|
|
|
|
binary data even on my slow workstation. |
|
234
|
|
|
|
|
|
|
|
|
235
|
|
|
|
|
|
|
=back |
|
236
|
|
|
|
|
|
|
|
|
237
|
|
|
|
|
|
|
I didn't know how to accomplish these tasks until I read something |
|
238
|
|
|
|
|
|
|
about XS. At least, it seemed as if it could solve my performance |
|
239
|
|
|
|
|
|
|
problems. However, writing a C parser in C isn't easier than it is |
|
240
|
|
|
|
|
|
|
in Perl. But writing a C preprocessor from scratch is even worse. |
|
241
|
|
|
|
|
|
|
|
|
242
|
|
|
|
|
|
|
Fortunately enough, after a few weeks of searching I found both, |
|
243
|
|
|
|
|
|
|
a lean, open-source C preprocessor library, and a reusable YACC |
|
244
|
|
|
|
|
|
|
grammar for ANSI-C. That was the beginning of the development of |
|
245
|
|
|
|
|
|
|
Convert::Binary::C in late 2001. |
|
246
|
|
|
|
|
|
|
|
|
247
|
|
|
|
|
|
|
Now, I'm successfully using the module in my embedded environment |
|
248
|
|
|
|
|
|
|
since long before it appeared on CPAN. From my point of view, it |
|
249
|
|
|
|
|
|
|
is exactly what I had in mind. It's fast, flexible, easy to use |
|
250
|
|
|
|
|
|
|
and portable. It doesn't require external programs or other Perl |
|
251
|
|
|
|
|
|
|
modules. |
|
252
|
|
|
|
|
|
|
|
|
253
|
|
|
|
|
|
|
=head2 About this document |
|
254
|
|
|
|
|
|
|
|
|
255
|
|
|
|
|
|
|
This document describes how to use Convert::Binary::C. A lot of |
|
256
|
|
|
|
|
|
|
different features are presented, and the example code sometimes |
|
257
|
|
|
|
|
|
|
uses Perl's more advanced language elements. If your experience |
|
258
|
|
|
|
|
|
|
with Perl is rather limited, you should know how to use Perl's |
|
259
|
|
|
|
|
|
|
very good documentation system. |
|
260
|
|
|
|
|
|
|
|
|
261
|
|
|
|
|
|
|
To look up one of the manpages, use the L<C<perldoc>|perldoc> command. |
|
262
|
|
|
|
|
|
|
For example, |
|
263
|
|
|
|
|
|
|
|
|
264
|
|
|
|
|
|
|
perldoc perl |
|
265
|
|
|
|
|
|
|
|
|
266
|
|
|
|
|
|
|
will show you Perl's main manpage. To look up a specific Perl |
|
267
|
|
|
|
|
|
|
function, use C<perldoc -f>: |
|
268
|
|
|
|
|
|
|
|
|
269
|
|
|
|
|
|
|
perldoc -f map |
|
270
|
|
|
|
|
|
|
|
|
271
|
|
|
|
|
|
|
gives you more information about the L<C<map>|perlfunc/"map"> function. |
|
272
|
|
|
|
|
|
|
You can also search the FAQ using C<perldoc -q>: |
|
273
|
|
|
|
|
|
|
|
|
274
|
|
|
|
|
|
|
perldoc -q array |
|
275
|
|
|
|
|
|
|
|
|
276
|
|
|
|
|
|
|
will give you everything you ever wanted to know about Perl |
|
277
|
|
|
|
|
|
|
arrays. But now, let's go on with some real stuff! |
|
278
|
|
|
|
|
|
|
|
|
279
|
|
|
|
|
|
|
=head2 Why use Convert::Binary::C? |
|
280
|
|
|
|
|
|
|
|
|
281
|
|
|
|
|
|
|
Say you want to pack (or unpack) data according to the following |
|
282
|
|
|
|
|
|
|
C structure: |
|
283
|
|
|
|
|
|
|
|
|
284
|
|
|
|
|
|
|
struct foo { |
|
285
|
|
|
|
|
|
|
char ary[3]; |
|
286
|
|
|
|
|
|
|
unsigned short baz; |
|
287
|
|
|
|
|
|
|
int bar; |
|
288
|
|
|
|
|
|
|
}; |
|
289
|
|
|
|
|
|
|
|
|
290
|
|
|
|
|
|
|
You could of course use |
|
291
|
|
|
|
|
|
|
Perl's L<C<pack>|perlfunc/"pack"> and L<C<unpack>|perlfunc/"unpack"> functions: |
|
292
|
|
|
|
|
|
|
|
|
293
|
|
|
|
|
|
|
@ary = (1, 2, 3); |
|
294
|
|
|
|
|
|
|
$baz = 40000; |
|
295
|
|
|
|
|
|
|
$bar = -4711; |
|
296
|
|
|
|
|
|
|
$binary = pack 'c3 S i', @ary, $baz, $bar; |
|
297
|
|
|
|
|
|
|
|
|
298
|
|
|
|
|
|
|
But this implies that the struct members are byte aligned. If |
|
299
|
|
|
|
|
|
|
they were long aligned (which is the default for most compilers), |
|
300
|
|
|
|
|
|
|
you'd have to write |
|
301
|
|
|
|
|
|
|
|
|
302
|
|
|
|
|
|
|
$binary = pack 'c3 x S x2 i', @ary, $baz, $bar; |
|
303
|
|
|
|
|
|
|
|
|
304
|
|
|
|
|
|
|
which doesn't really increase readability. |
|
305
|
|
|
|
|
|
|
|
|
306
|
|
|
|
|
|
|
Now imagine that you need to pack the data for a completely |
|
307
|
|
|
|
|
|
|
different architecture with different byte order. You would |
|
308
|
|
|
|
|
|
|
look into the L<C<pack>|perlfunc/"pack"> manpage again and |
|
309
|
|
|
|
|
|
|
perhaps come up with this: |
|
310
|
|
|
|
|
|
|
|
|
311
|
|
|
|
|
|
|
$binary = pack 'c3 x n x2 N', @ary, $baz, $bar; |
|
312
|
|
|
|
|
|
|
|
|
313
|
|
|
|
|
|
|
However, if you try to unpack C<$foo> again, your signed values |
|
314
|
|
|
|
|
|
|
have turned into unsigned ones. |
|
315
|
|
|
|
|
|
|
|
|
316
|
|
|
|
|
|
|
All this can still be managed with Perl. But imagine your |
|
317
|
|
|
|
|
|
|
structures get more complex? Imagine you need to support |
|
318
|
|
|
|
|
|
|
different platforms? Imagine you need to make changes to |
|
319
|
|
|
|
|
|
|
the structures? You'll not only have to change the C source |
|
320
|
|
|
|
|
|
|
but also dozens of L<C<pack>|perlfunc/"pack"> strings in |
|
321
|
|
|
|
|
|
|
your Perl code. This is no fun. And Perl should be fun. |
|
322
|
|
|
|
|
|
|
|
|
323
|
|
|
|
|
|
|
Now, wouldn't it be great if you could just read in the C |
|
324
|
|
|
|
|
|
|
source you've already written and use all the types defined |
|
325
|
|
|
|
|
|
|
there for packing and unpacking? That's what Convert::Binary::C |
|
326
|
|
|
|
|
|
|
does. |
|
327
|
|
|
|
|
|
|
|
|
328
|
|
|
|
|
|
|
=head2 Creating a Convert::Binary::C object |
|
329
|
|
|
|
|
|
|
|
|
330
|
|
|
|
|
|
|
To use Convert::Binary::C just say |
|
331
|
|
|
|
|
|
|
|
|
332
|
|
|
|
|
|
|
use Convert::Binary::C; |
|
333
|
|
|
|
|
|
|
|
|
334
|
|
|
|
|
|
|
to load the module. Its interface is completely object |
|
335
|
|
|
|
|
|
|
oriented, so it doesn't export any functions. |
|
336
|
|
|
|
|
|
|
|
|
337
|
|
|
|
|
|
|
Next, you need to create a new Convert::Binary::C object. This |
|
338
|
|
|
|
|
|
|
can be done by either |
|
339
|
|
|
|
|
|
|
|
|
340
|
|
|
|
|
|
|
$c = Convert::Binary::C->new; |
|
341
|
|
|
|
|
|
|
|
|
342
|
|
|
|
|
|
|
or |
|
343
|
|
|
|
|
|
|
|
|
344
|
|
|
|
|
|
|
$c = new Convert::Binary::C; |
|
345
|
|
|
|
|
|
|
|
|
346
|
|
|
|
|
|
|
You can optionally pass configuration options to |
|
347
|
|
|
|
|
|
|
the L<constructor|/"new"> as described in the next section. |
|
348
|
|
|
|
|
|
|
|
|
349
|
|
|
|
|
|
|
=head2 Configuring the object |
|
350
|
|
|
|
|
|
|
|
|
351
|
|
|
|
|
|
|
To configure a Convert::Binary::C object, you can either call |
|
352
|
|
|
|
|
|
|
the L<C<configure>|/"configure"> method or directly pass the configuration |
|
353
|
|
|
|
|
|
|
options to the L<constructor|/"new">. If you want to change byte order |
|
354
|
|
|
|
|
|
|
and alignment, you can use |
|
355
|
|
|
|
|
|
|
|
|
356
|
|
|
|
|
|
|
$c->configure(ByteOrder => 'LittleEndian', |
|
357
|
|
|
|
|
|
|
Alignment => 2); |
|
358
|
|
|
|
|
|
|
|
|
359
|
|
|
|
|
|
|
or you can change the construction code to |
|
360
|
|
|
|
|
|
|
|
|
361
|
|
|
|
|
|
|
$c = new Convert::Binary::C ByteOrder => 'LittleEndian', |
|
362
|
|
|
|
|
|
|
Alignment => 2; |
|
363
|
|
|
|
|
|
|
|
|
364
|
|
|
|
|
|
|
Either way, the object will now know that it should use |
|
365
|
|
|
|
|
|
|
little endian (Intel) byte order and 2-byte struct member |
|
366
|
|
|
|
|
|
|
alignment for packing and unpacking. |
|
367
|
|
|
|
|
|
|
|
|
368
|
|
|
|
|
|
|
Alternatively, you can use the option names as names of |
|
369
|
|
|
|
|
|
|
methods to configure the object, like: |
|
370
|
|
|
|
|
|
|
|
|
371
|
|
|
|
|
|
|
$c->ByteOrder('LittleEndian'); |
|
372
|
|
|
|
|
|
|
|
|
373
|
|
|
|
|
|
|
You can also retrieve information about the current |
|
374
|
|
|
|
|
|
|
configuration of a Convert::Binary::C object. For details, |
|
375
|
|
|
|
|
|
|
see the section about the L<C<configure>|/"configure"> method. |
|
376
|
|
|
|
|
|
|
|
|
377
|
|
|
|
|
|
|
=head2 Parsing C code |
|
378
|
|
|
|
|
|
|
|
|
379
|
|
|
|
|
|
|
Convert::Binary::C allows two ways of parsing C source. Either |
|
380
|
|
|
|
|
|
|
by parsing external C header or C source files: |
|
381
|
|
|
|
|
|
|
|
|
382
|
|
|
|
|
|
|
$c->parse_file('header.h'); |
|
383
|
|
|
|
|
|
|
|
|
384
|
|
|
|
|
|
|
Or by parsing C code embedded in your script: |
|
385
|
|
|
|
|
|
|
|
|
386
|
|
|
|
|
|
|
$c->parse(<<'CCODE'); |
|
387
|
|
|
|
|
|
|
struct foo { |
|
388
|
|
|
|
|
|
|
char ary[3]; |
|
389
|
|
|
|
|
|
|
unsigned short baz; |
|
390
|
|
|
|
|
|
|
int bar; |
|
391
|
|
|
|
|
|
|
}; |
|
392
|
|
|
|
|
|
|
CCODE |
|
393
|
|
|
|
|
|
|
|
|
394
|
|
|
|
|
|
|
Now the object C<$c> will know everything about C<struct foo>. |
|
395
|
|
|
|
|
|
|
The example above uses a so-called here-document. It allows to |
|
396
|
|
|
|
|
|
|
easily embed multi-line strings in your code. You can find more |
|
397
|
|
|
|
|
|
|
about here-documents in L<perldata> or L<perlop>. |
|
398
|
|
|
|
|
|
|
|
|
399
|
|
|
|
|
|
|
Since the L<C<parse>|/"parse"> and L<C<parse_file>|/"parse_file"> methods |
|
400
|
|
|
|
|
|
|
throw an exception when a parse error occurs, you usually want to catch |
|
401
|
|
|
|
|
|
|
these in an C<eval> block: |
|
402
|
|
|
|
|
|
|
|
|
403
|
|
|
|
|
|
|
eval { $c->parse_file('header.h') }; |
|
404
|
|
|
|
|
|
|
if ($@) { |
|
405
|
|
|
|
|
|
|
# handle error appropriately |
|
406
|
|
|
|
|
|
|
} |
|
407
|
|
|
|
|
|
|
|
|
408
|
|
|
|
|
|
|
Perl's special C<$@> variable will contain an empty string (which |
|
409
|
|
|
|
|
|
|
evaluates to a false value in boolean context) on success or |
|
410
|
|
|
|
|
|
|
an error string on failure. |
|
411
|
|
|
|
|
|
|
|
|
412
|
|
|
|
|
|
|
As another feature, L<C<parse>|/"parse"> and L<C<parse_file>|/"parse_file"> return |
|
413
|
|
|
|
|
|
|
a reference to their object on success, just like L<C<configure>|/"configure"> does |
|
414
|
|
|
|
|
|
|
when you're configuring the object. This will allow you to write constructs |
|
415
|
|
|
|
|
|
|
like this: |
|
416
|
|
|
|
|
|
|
|
|
417
|
|
|
|
|
|
|
my $c = eval { |
|
418
|
|
|
|
|
|
|
Convert::Binary::C->new(Include => ['/usr/include']) |
|
419
|
|
|
|
|
|
|
->parse_file('header.h') |
|
420
|
|
|
|
|
|
|
}; |
|
421
|
|
|
|
|
|
|
if ($@) { |
|
422
|
|
|
|
|
|
|
# handle error appropriately |
|
423
|
|
|
|
|
|
|
} |
|
424
|
|
|
|
|
|
|
|
|
425
|
|
|
|
|
|
|
=head2 Packing and unpacking |
|
426
|
|
|
|
|
|
|
|
|
427
|
|
|
|
|
|
|
Convert::Binary::C has two methods, L<C<pack>|/"pack"> and L<C<unpack>|/"unpack">, |
|
428
|
|
|
|
|
|
|
that act similar to the functions of same denominator in Perl. |
|
429
|
|
|
|
|
|
|
To perform the packing described in the example above, |
|
430
|
|
|
|
|
|
|
you could write: |
|
431
|
|
|
|
|
|
|
|
|
432
|
|
|
|
|
|
|
$data = { |
|
433
|
|
|
|
|
|
|
ary => [1, 2, 3], |
|
434
|
|
|
|
|
|
|
baz => 40000, |
|
435
|
|
|
|
|
|
|
bar => -4711, |
|
436
|
|
|
|
|
|
|
}; |
|
437
|
|
|
|
|
|
|
$binary = $c->pack('foo', $data); |
|
438
|
|
|
|
|
|
|
|
|
439
|
|
|
|
|
|
|
Unpacking will work exactly the same way, just that |
|
440
|
|
|
|
|
|
|
the L<C<unpack>|/"unpack"> method will take a byte string as its input |
|
441
|
|
|
|
|
|
|
and will return a reference to a (possibly very complex) |
|
442
|
|
|
|
|
|
|
Perl data structure. |
|
443
|
|
|
|
|
|
|
|
|
444
|
|
|
|
|
|
|
$binary = get_data_from_memory(); |
|
445
|
|
|
|
|
|
|
$data = $c->unpack('foo', $binary); |
|
446
|
|
|
|
|
|
|
|
|
447
|
|
|
|
|
|
|
You can now easily access all of the values: |
|
448
|
|
|
|
|
|
|
|
|
449
|
|
|
|
|
|
|
print "foo.ary[1] = $data->{ary}[1]\n"; |
|
450
|
|
|
|
|
|
|
|
|
451
|
|
|
|
|
|
|
Or you can even more conveniently use |
|
452
|
|
|
|
|
|
|
the L<Data::Dumper|Data::Dumper> module: |
|
453
|
|
|
|
|
|
|
|
|
454
|
|
|
|
|
|
|
use Data::Dumper; |
|
455
|
|
|
|
|
|
|
print Dumper($data); |
|
456
|
|
|
|
|
|
|
|
|
457
|
|
|
|
|
|
|
The output would look something like this: |
|
458
|
|
|
|
|
|
|
|
|
459
|
|
|
|
|
|
|
$VAR1 = { |
|
460
|
|
|
|
|
|
|
'bar' => -271, |
|
461
|
|
|
|
|
|
|
'baz' => 5000, |
|
462
|
|
|
|
|
|
|
'ary' => [ |
|
463
|
|
|
|
|
|
|
42, |
|
464
|
|
|
|
|
|
|
48, |
|
465
|
|
|
|
|
|
|
100 |
|
466
|
|
|
|
|
|
|
] |
|
467
|
|
|
|
|
|
|
}; |
|
468
|
|
|
|
|
|
|
|
|
469
|
|
|
|
|
|
|
=head2 Preprocessor configuration |
|
470
|
|
|
|
|
|
|
|
|
471
|
|
|
|
|
|
|
Convert::Binary::C uses Thomas Pornin's C<ucpp> as an internal |
|
472
|
|
|
|
|
|
|
C preprocessor. It is compliant to ISO-C99, so you don't have |
|
473
|
|
|
|
|
|
|
to worry about using even weird preprocessor constructs in |
|
474
|
|
|
|
|
|
|
your code. |
|
475
|
|
|
|
|
|
|
|
|
476
|
|
|
|
|
|
|
If your C source contains includes or depends upon preprocessor |
|
477
|
|
|
|
|
|
|
defines, you may need to configure the internal preprocessor. |
|
478
|
|
|
|
|
|
|
Use the C<Include> and C<Define> configuration options for that: |
|
479
|
|
|
|
|
|
|
|
|
480
|
|
|
|
|
|
|
$c->configure(Include => ['/usr/include', |
|
481
|
|
|
|
|
|
|
'/home/mhx/include'], |
|
482
|
|
|
|
|
|
|
Define => [qw( NDEBUG FOO=42 )]); |
|
483
|
|
|
|
|
|
|
|
|
484
|
|
|
|
|
|
|
If your code uses system includes, it is most likely |
|
485
|
|
|
|
|
|
|
that you will need to define the symbols that are usually |
|
486
|
|
|
|
|
|
|
defined by the compiler. |
|
487
|
|
|
|
|
|
|
|
|
488
|
|
|
|
|
|
|
On some operating systems, the system includes require the |
|
489
|
|
|
|
|
|
|
preprocessor to predefine a certain set of assertions. |
|
490
|
|
|
|
|
|
|
Assertions are supported by C<ucpp>, and you can define them |
|
491
|
|
|
|
|
|
|
either in the source code using C<#assert> or as a property |
|
492
|
|
|
|
|
|
|
of the Convert::Binary::C object using C<Assert>: |
|
493
|
|
|
|
|
|
|
|
|
494
|
|
|
|
|
|
|
$c->configure(Assert => ['predicate(answer)']); |
|
495
|
|
|
|
|
|
|
|
|
496
|
|
|
|
|
|
|
Information about defined macros can be retrieved from the |
|
497
|
|
|
|
|
|
|
preprocessor as long as its configuration isn't changed. The |
|
498
|
|
|
|
|
|
|
preprocessor is implicitly reset if you change one of the |
|
499
|
|
|
|
|
|
|
following configuration options: |
|
500
|
|
|
|
|
|
|
|
|
501
|
|
|
|
|
|
|
Include |
|
502
|
|
|
|
|
|
|
Define |
|
503
|
|
|
|
|
|
|
Assert |
|
504
|
|
|
|
|
|
|
HasCPPComments |
|
505
|
|
|
|
|
|
|
HasMacroVAARGS |
|
506
|
|
|
|
|
|
|
|
|
507
|
|
|
|
|
|
|
=head2 Supported pragma directives |
|
508
|
|
|
|
|
|
|
|
|
509
|
|
|
|
|
|
|
Convert::Binary::C supports the C<pack> pragma to locally override |
|
510
|
|
|
|
|
|
|
struct member alignment. The supported syntax is as follows: |
|
511
|
|
|
|
|
|
|
|
|
512
|
|
|
|
|
|
|
=over 4 |
|
513
|
|
|
|
|
|
|
|
|
514
|
|
|
|
|
|
|
=item #pragma pack( ALIGN ) |
|
515
|
|
|
|
|
|
|
|
|
516
|
|
|
|
|
|
|
Sets the new alignment to ALIGN. |
|
517
|
|
|
|
|
|
|
|
|
518
|
|
|
|
|
|
|
=item #pragma pack |
|
519
|
|
|
|
|
|
|
|
|
520
|
|
|
|
|
|
|
Resets the alignment to its original value. |
|
521
|
|
|
|
|
|
|
|
|
522
|
|
|
|
|
|
|
=item #pragma pack( push, ALIGN ) |
|
523
|
|
|
|
|
|
|
|
|
524
|
|
|
|
|
|
|
Saves the current alignment on a stack and sets the new |
|
525
|
|
|
|
|
|
|
alignment to ALIGN. |
|
526
|
|
|
|
|
|
|
|
|
527
|
|
|
|
|
|
|
=item #pragma pack( pop ) |
|
528
|
|
|
|
|
|
|
|
|
529
|
|
|
|
|
|
|
Restores the alignment to the last value saved on the |
|
530
|
|
|
|
|
|
|
stack. |
|
531
|
|
|
|
|
|
|
|
|
532
|
|
|
|
|
|
|
=back |
|
533
|
|
|
|
|
|
|
|
|
534
|
|
|
|
|
|
|
/* Example assumes sizeof( short ) == 2, sizeof( long ) == 4. */ |
|
535
|
|
|
|
|
|
|
|
|
536
|
|
|
|
|
|
|
#pragma pack(1) |
|
537
|
|
|
|
|
|
|
|
|
538
|
|
|
|
|
|
|
struct nopad { |
|
539
|
|
|
|
|
|
|
char a; /* no padding bytes between 'a' and 'b' */ |
|
540
|
|
|
|
|
|
|
long b; |
|
541
|
|
|
|
|
|
|
}; |
|
542
|
|
|
|
|
|
|
|
|
543
|
|
|
|
|
|
|
#pragma pack /* reset to "native" alignment */ |
|
544
|
|
|
|
|
|
|
|
|
545
|
|
|
|
|
|
|
#pragma pack( push, 2 ) |
|
546
|
|
|
|
|
|
|
|
|
547
|
|
|
|
|
|
|
struct pad { |
|
548
|
|
|
|
|
|
|
char a; /* one padding byte between 'a' and 'b' */ |
|
549
|
|
|
|
|
|
|
long b; |
|
550
|
|
|
|
|
|
|
|
|
551
|
|
|
|
|
|
|
#pragma pack( push, 1 ) |
|
552
|
|
|
|
|
|
|
|
|
553
|
|
|
|
|
|
|
struct { |
|
554
|
|
|
|
|
|
|
char c; /* no padding between 'c' and 'd' */ |
|
555
|
|
|
|
|
|
|
short d; |
|
556
|
|
|
|
|
|
|
} e; /* sizeof( e ) == 3 */ |
|
557
|
|
|
|
|
|
|
|
|
558
|
|
|
|
|
|
|
#pragma pack( pop ); /* back to pack( 2 ) */ |
|
559
|
|
|
|
|
|
|
|
|
560
|
|
|
|
|
|
|
long f; /* one padding byte between 'e' and 'f' */ |
|
561
|
|
|
|
|
|
|
}; |
|
562
|
|
|
|
|
|
|
|
|
563
|
|
|
|
|
|
|
#pragma pack( pop ); /* back to "native" */ |
|
564
|
|
|
|
|
|
|
|
|
565
|
|
|
|
|
|
|
The C<pack> pragma as it is currently implemented only affects |
|
566
|
|
|
|
|
|
|
the I<maximum> struct member alignment. There are compilers |
|
567
|
|
|
|
|
|
|
that also allow to specify the I<minimum> struct member |
|
568
|
|
|
|
|
|
|
alignment. This is not supported by Convert::Binary::C. |
|
569
|
|
|
|
|
|
|
|
|
570
|
|
|
|
|
|
|
=head2 Automatic configuration using C<ccconfig> |
|
571
|
|
|
|
|
|
|
|
|
572
|
|
|
|
|
|
|
As there are over 20 different configuration options, setting |
|
573
|
|
|
|
|
|
|
all of them correctly can be a lengthy and tedious task. |
|
574
|
|
|
|
|
|
|
|
|
575
|
|
|
|
|
|
|
The L<C<ccconfig>|ccconfig> script, which is bundled with this |
|
576
|
|
|
|
|
|
|
module, aims at automatically determining the correct compiler |
|
577
|
|
|
|
|
|
|
configuration by testing the compiler executable. It works for |
|
578
|
|
|
|
|
|
|
both, native and cross compilers. |
|
579
|
|
|
|
|
|
|
|
|
580
|
|
|
|
|
|
|
=head1 UNDERSTANDING TYPES |
|
581
|
|
|
|
|
|
|
|
|
582
|
|
|
|
|
|
|
This section covers one of the fundamental features of |
|
583
|
|
|
|
|
|
|
Convert::Binary::C. It's how I<type expressions>, referred to |
|
584
|
|
|
|
|
|
|
as TYPEs in the L<method reference|/"METHODS">, are handled |
|
585
|
|
|
|
|
|
|
by the module. |
|
586
|
|
|
|
|
|
|
|
|
587
|
|
|
|
|
|
|
Many of the methods, |
|
588
|
|
|
|
|
|
|
namely L<C<pack>|/"pack">, L<C<unpack>|/"unpack">, L<C<sizeof>|/"sizeof">, L<C<typeof>|/"typeof">, L<C<member>|/"member">, L<C<offsetof>|/"offsetof">, L<C<def>|/"def">, L<C<initializer>|/"initializer"> and L<C<tag>|/"tag">, |
|
589
|
|
|
|
|
|
|
are passed a TYPE to operate on as their first argument. |
|
590
|
|
|
|
|
|
|
|
|
591
|
|
|
|
|
|
|
=head2 Standard Types |
|
592
|
|
|
|
|
|
|
|
|
593
|
|
|
|
|
|
|
These are trivial. Standard types are simply enum names, struct |
|
594
|
|
|
|
|
|
|
names, union names, or typedefs. Almost every method that wants |
|
595
|
|
|
|
|
|
|
a TYPE will accept a standard type. |
|
596
|
|
|
|
|
|
|
|
|
597
|
|
|
|
|
|
|
For enums, structs and unions, the prefixes C<enum>, C<struct> and C<union> are |
|
598
|
|
|
|
|
|
|
optional. However, if a typedef with the same name exists, like in |
|
599
|
|
|
|
|
|
|
|
|
600
|
|
|
|
|
|
|
struct foo { |
|
601
|
|
|
|
|
|
|
int bar; |
|
602
|
|
|
|
|
|
|
}; |
|
603
|
|
|
|
|
|
|
|
|
604
|
|
|
|
|
|
|
typedef int foo; |
|
605
|
|
|
|
|
|
|
|
|
606
|
|
|
|
|
|
|
you will have to use the prefix to distinguish between the |
|
607
|
|
|
|
|
|
|
struct and the typedef. Otherwise, a typedef is always given |
|
608
|
|
|
|
|
|
|
preference. |
|
609
|
|
|
|
|
|
|
|
|
610
|
|
|
|
|
|
|
=head2 Basic Types |
|
611
|
|
|
|
|
|
|
|
|
612
|
|
|
|
|
|
|
Basic types, or atomic types, are C<int> or C<char>, for example. |
|
613
|
|
|
|
|
|
|
It's possible to use these basic types without having parsed any |
|
614
|
|
|
|
|
|
|
code. You can simply do |
|
615
|
|
|
|
|
|
|
|
|
616
|
|
|
|
|
|
|
$c = new Convert::Binary::C; |
|
617
|
|
|
|
|
|
|
$size = $c->sizeof('unsigned long'); |
|
618
|
|
|
|
|
|
|
$data = $c->pack('short int', 42); |
|
619
|
|
|
|
|
|
|
|
|
620
|
|
|
|
|
|
|
Even though the above works fine, it is not possible to define |
|
621
|
|
|
|
|
|
|
more complex types on the fly, so |
|
622
|
|
|
|
|
|
|
|
|
623
|
|
|
|
|
|
|
$size = $c->sizeof('struct { int a, b; }'); |
|
624
|
|
|
|
|
|
|
|
|
625
|
|
|
|
|
|
|
will result in an error. |
|
626
|
|
|
|
|
|
|
|
|
627
|
|
|
|
|
|
|
Basic types are not supported by all methods. For example, it makes |
|
628
|
|
|
|
|
|
|
no sense to use L<C<member>|/"member"> or L<C<offsetof>|/"offsetof"> on |
|
629
|
|
|
|
|
|
|
a basic type. Using L<C<typeof>|/"typeof"> isn't very useful, but |
|
630
|
|
|
|
|
|
|
supported. |
|
631
|
|
|
|
|
|
|
|
|
632
|
|
|
|
|
|
|
=head2 Member Expressions |
|
633
|
|
|
|
|
|
|
|
|
634
|
|
|
|
|
|
|
This is by far the most complex part, depending on the complexity of |
|
635
|
|
|
|
|
|
|
your data structures. Any L<standard type|/"Standard Types"> that |
|
636
|
|
|
|
|
|
|
defines a compound or an array may be followed by a member expression |
|
637
|
|
|
|
|
|
|
to select only a certain part of the data type. Say you have parsed the |
|
638
|
|
|
|
|
|
|
following C code: |
|
639
|
|
|
|
|
|
|
|
|
640
|
|
|
|
|
|
|
struct foo { |
|
641
|
|
|
|
|
|
|
long type; |
|
642
|
|
|
|
|
|
|
struct { |
|
643
|
|
|
|
|
|
|
short x, y; |
|
644
|
|
|
|
|
|
|
} array[20]; |
|
645
|
|
|
|
|
|
|
}; |
|
646
|
|
|
|
|
|
|
|
|
647
|
|
|
|
|
|
|
typedef struct foo matrix[8][8]; |
|
648
|
|
|
|
|
|
|
|
|
649
|
|
|
|
|
|
|
You may want to know the size of the C<array> member of C<struct foo>. |
|
650
|
|
|
|
|
|
|
This is quite easy: |
|
651
|
|
|
|
|
|
|
|
|
652
|
|
|
|
|
|
|
print $c->sizeof('foo.array'), " bytes"; |
|
653
|
|
|
|
|
|
|
|
|
654
|
|
|
|
|
|
|
will print |
|
655
|
|
|
|
|
|
|
|
|
656
|
|
|
|
|
|
|
80 bytes |
|
657
|
|
|
|
|
|
|
|
|
658
|
|
|
|
|
|
|
depending of course on the C<ShortSize> you configured. |
|
659
|
|
|
|
|
|
|
|
|
660
|
|
|
|
|
|
|
If you wanted to unpack only a single column of C<matrix>, that's |
|
661
|
|
|
|
|
|
|
easy as well (and of course it doesn't matter which index you use): |
|
662
|
|
|
|
|
|
|
|
|
663
|
|
|
|
|
|
|
$column = $c->unpack('matrix[2]', $data); |
|
664
|
|
|
|
|
|
|
|
|
665
|
|
|
|
|
|
|
Just like in C, it is possible to use out-of-bounds array indices. |
|
666
|
|
|
|
|
|
|
This means that, for example, despite C<array> is declared to have |
|
667
|
|
|
|
|
|
|
20 elements, the following code |
|
668
|
|
|
|
|
|
|
|
|
669
|
|
|
|
|
|
|
$size = $c->sizeof('foo.array[4711]'); |
|
670
|
|
|
|
|
|
|
$offset = $c->offsetof('foo', 'array[-13]'); |
|
671
|
|
|
|
|
|
|
|
|
672
|
|
|
|
|
|
|
is perfectly valid and will result in: |
|
673
|
|
|
|
|
|
|
|
|
674
|
|
|
|
|
|
|
$size = 4 |
|
675
|
|
|
|
|
|
|
$offset = -48 |
|
676
|
|
|
|
|
|
|
|
|
677
|
|
|
|
|
|
|
Member expressions can be arbitrarily complex: |
|
678
|
|
|
|
|
|
|
|
|
679
|
|
|
|
|
|
|
$type = $c->typeof('matrix[2][3].array[7].y'); |
|
680
|
|
|
|
|
|
|
print "the type is $type"; |
|
681
|
|
|
|
|
|
|
|
|
682
|
|
|
|
|
|
|
will, for example, print |
|
683
|
|
|
|
|
|
|
|
|
684
|
|
|
|
|
|
|
the type is short |
|
685
|
|
|
|
|
|
|
|
|
686
|
|
|
|
|
|
|
Member expressions are also used as the second argument |
|
687
|
|
|
|
|
|
|
to L<C<offsetof>|/"offsetof">. |
|
688
|
|
|
|
|
|
|
|
|
689
|
|
|
|
|
|
|
=head2 Offsets |
|
690
|
|
|
|
|
|
|
|
|
691
|
|
|
|
|
|
|
Members returned by the L<C<member>|/"member"> method have an optional |
|
692
|
|
|
|
|
|
|
offset suffix to indicate that the given offset doesn't point to the |
|
693
|
|
|
|
|
|
|
start of that member. For example, |
|
694
|
|
|
|
|
|
|
|
|
695
|
|
|
|
|
|
|
$member = $c->member('matrix', 1431); |
|
696
|
|
|
|
|
|
|
print $member; |
|
697
|
|
|
|
|
|
|
|
|
698
|
|
|
|
|
|
|
will print |
|
699
|
|
|
|
|
|
|
|
|
700
|
|
|
|
|
|
|
[2][1].type+3 |
|
701
|
|
|
|
|
|
|
|
|
702
|
|
|
|
|
|
|
If you would use this as a member expression, like in |
|
703
|
|
|
|
|
|
|
|
|
704
|
|
|
|
|
|
|
$size = $c->sizeof("matrix $member"); |
|
705
|
|
|
|
|
|
|
|
|
706
|
|
|
|
|
|
|
the offset suffix will simply be ignored. Actually, it will be |
|
707
|
|
|
|
|
|
|
ignored for all methods if it's used in the first argument. |
|
708
|
|
|
|
|
|
|
|
|
709
|
|
|
|
|
|
|
When used in the second argument to L<C<offsetof>|/"offsetof">, |
|
710
|
|
|
|
|
|
|
it will usually do what you mean, i. e. the offset suffix, if |
|
711
|
|
|
|
|
|
|
present, will be considered when determining the offset. This |
|
712
|
|
|
|
|
|
|
behaviour ensures that |
|
713
|
|
|
|
|
|
|
|
|
714
|
|
|
|
|
|
|
$member = $c->member('foo', 43); |
|
715
|
|
|
|
|
|
|
$offset = $c->offsetof('foo', $member); |
|
716
|
|
|
|
|
|
|
print "'$member' is located at offset $offset of struct foo"; |
|
717
|
|
|
|
|
|
|
|
|
718
|
|
|
|
|
|
|
will always correctly set C<$offset>: |
|
719
|
|
|
|
|
|
|
|
|
720
|
|
|
|
|
|
|
'.array[9].y+1' is located at offset 43 of struct foo |
|
721
|
|
|
|
|
|
|
|
|
722
|
|
|
|
|
|
|
If this is not what you mean, e.g. because you want to know the |
|
723
|
|
|
|
|
|
|
offset where the member returned by L<C<member>|/"member"> starts, |
|
724
|
|
|
|
|
|
|
you just have to remove the suffix: |
|
725
|
|
|
|
|
|
|
|
|
726
|
|
|
|
|
|
|
$member =~ s/\+\d+$//; |
|
727
|
|
|
|
|
|
|
$offset = $c->offsetof('foo', $member); |
|
728
|
|
|
|
|
|
|
print "'$member' starts at offset $offset of struct foo"; |
|
729
|
|
|
|
|
|
|
|
|
730
|
|
|
|
|
|
|
This would then print: |
|
731
|
|
|
|
|
|
|
|
|
732
|
|
|
|
|
|
|
'.array[9].y' starts at offset 42 of struct foo |
|
733
|
|
|
|
|
|
|
|
|
734
|
|
|
|
|
|
|
=head1 USING TAGS |
|
735
|
|
|
|
|
|
|
|
|
736
|
|
|
|
|
|
|
In a nutshell, tags are properties that you can attach to types. |
|
737
|
|
|
|
|
|
|
|
|
738
|
|
|
|
|
|
|
You can add tags to types using the L<C<tag>|/"tag"> method, |
|
739
|
|
|
|
|
|
|
and remove them using L<C<tag>|/"tag"> or L<C<untag>|/"untag">, |
|
740
|
|
|
|
|
|
|
for example: |
|
741
|
|
|
|
|
|
|
|
|
742
|
|
|
|
|
|
|
# Attach 'Format' and 'Hooks' tags |
|
743
|
|
|
|
|
|
|
$c->tag('type', Format => 'String', Hooks => { pack => \&rout }); |
|
744
|
|
|
|
|
|
|
|
|
745
|
|
|
|
|
|
|
$c->untag('type', 'Format'); # Remove only 'Format' tag |
|
746
|
|
|
|
|
|
|
$c->untag('type'); # Remove all tags |
|
747
|
|
|
|
|
|
|
|
|
748
|
|
|
|
|
|
|
You can also use L<C<tag>|/"tag"> to see which tags are |
|
749
|
|
|
|
|
|
|
attached to a type, for example: |
|
750
|
|
|
|
|
|
|
|
|
751
|
|
|
|
|
|
|
$tags = $c->tag('type'); |
|
752
|
|
|
|
|
|
|
|
|
753
|
|
|
|
|
|
|
This would give you: |
|
754
|
|
|
|
|
|
|
|
|
755
|
|
|
|
|
|
|
$tags = { |
|
756
|
|
|
|
|
|
|
'Hooks' => { |
|
757
|
|
|
|
|
|
|
'pack' => \&rout |
|
758
|
|
|
|
|
|
|
}, |
|
759
|
|
|
|
|
|
|
'Format' => 'String' |
|
760
|
|
|
|
|
|
|
}; |
|
761
|
|
|
|
|
|
|
|
|
762
|
|
|
|
|
|
|
Currently, there are only a couple of different tags that |
|
763
|
|
|
|
|
|
|
influence the way data is packed and unpacked. There are |
|
764
|
|
|
|
|
|
|
probably more tags to come in the future. |
|
765
|
|
|
|
|
|
|
|
|
766
|
|
|
|
|
|
|
=head2 The Format Tag |
|
767
|
|
|
|
|
|
|
|
|
768
|
|
|
|
|
|
|
One of the tags currently available is the C<Format> tag. |
|
769
|
|
|
|
|
|
|
Using this tag, you can tell a Convert::Binary::C object to |
|
770
|
|
|
|
|
|
|
pack and unpack a certain data type in a special way. |
|
771
|
|
|
|
|
|
|
|
|
772
|
|
|
|
|
|
|
For example, if you have a (fixed length) string type |
|
773
|
|
|
|
|
|
|
|
|
774
|
|
|
|
|
|
|
typedef char str_type[40]; |
|
775
|
|
|
|
|
|
|
|
|
776
|
|
|
|
|
|
|
this type would, by default, be unpacked as an array |
|
777
|
|
|
|
|
|
|
of C<char>s. That's because it B<is> only an array |
|
778
|
|
|
|
|
|
|
of C<char>s, and Convert::Binary::C doesn't know it is |
|
779
|
|
|
|
|
|
|
actually used as a string. |
|
780
|
|
|
|
|
|
|
|
|
781
|
|
|
|
|
|
|
But you can tell Convert::Binary::C that C<str_type> is |
|
782
|
|
|
|
|
|
|
a C string using the C<Format> tag: |
|
783
|
|
|
|
|
|
|
|
|
784
|
|
|
|
|
|
|
$c->tag('str_type', Format => 'String'); |
|
785
|
|
|
|
|
|
|
|
|
786
|
|
|
|
|
|
|
This will make L<C<unpack>|/"unpack"> (and of course |
|
787
|
|
|
|
|
|
|
also L<C<pack>|/"pack">) treat the binary data like a |
|
788
|
|
|
|
|
|
|
null-terminated C string: |
|
789
|
|
|
|
|
|
|
|
|
790
|
|
|
|
|
|
|
$binary = "Hello World!\n\0 this is just some dummy data"; |
|
791
|
|
|
|
|
|
|
$hello = $c->unpack('str_type', $binary); |
|
792
|
|
|
|
|
|
|
print $hello; |
|
793
|
|
|
|
|
|
|
|
|
794
|
|
|
|
|
|
|
would thusly print: |
|
795
|
|
|
|
|
|
|
|
|
796
|
|
|
|
|
|
|
Hello World! |
|
797
|
|
|
|
|
|
|
|
|
798
|
|
|
|
|
|
|
Of course, this also works the other way round: |
|
799
|
|
|
|
|
|
|
|
|
800
|
|
|
|
|
|
|
use Data::Hexdumper; |
|
801
|
|
|
|
|
|
|
|
|
802
|
|
|
|
|
|
|
$binary = $c->pack('str_type', "Just another C::B::C hacker"); |
|
803
|
|
|
|
|
|
|
print hexdump(data => $binary); |
|
804
|
|
|
|
|
|
|
|
|
805
|
|
|
|
|
|
|
would print: |
|
806
|
|
|
|
|
|
|
|
|
807
|
|
|
|
|
|
|
0x0000 : 4A 75 73 74 20 61 6E 6F 74 68 65 72 20 43 3A 3A : Just.another.C:: |
|
808
|
|
|
|
|
|
|
0x0010 : 42 3A 3A 43 20 68 61 63 6B 65 72 00 00 00 00 00 : B::C.hacker..... |
|
809
|
|
|
|
|
|
|
0x0020 : 00 00 00 00 00 00 00 00 : ........ |
|
810
|
|
|
|
|
|
|
|
|
811
|
|
|
|
|
|
|
If you want Convert::Binary::C to not interpret the binary |
|
812
|
|
|
|
|
|
|
data at all, you can set the C<Format> tag to C<Binary>. |
|
813
|
|
|
|
|
|
|
This might not be seem very useful, |
|
814
|
|
|
|
|
|
|
as L<C<pack>|/"pack"> and L<C<unpack>|/"unpack"> would |
|
815
|
|
|
|
|
|
|
just pass through the unmodified binary data. |
|
816
|
|
|
|
|
|
|
But you can tag not only whole types, but also compound |
|
817
|
|
|
|
|
|
|
members. For example |
|
818
|
|
|
|
|
|
|
|
|
819
|
|
|
|
|
|
|
$c->parse(<<ENDC); |
|
820
|
|
|
|
|
|
|
struct packet { |
|
821
|
|
|
|
|
|
|
unsigned short header; |
|
822
|
|
|
|
|
|
|
unsigned short flags; |
|
823
|
|
|
|
|
|
|
unsigned char payload[28]; |
|
824
|
|
|
|
|
|
|
}; |
|
825
|
|
|
|
|
|
|
ENDC |
|
826
|
|
|
|
|
|
|
|
|
827
|
|
|
|
|
|
|
$c->tag('packet.payload', Format => 'Binary'); |
|
828
|
|
|
|
|
|
|
|
|
829
|
|
|
|
|
|
|
would allow you to write: |
|
830
|
|
|
|
|
|
|
|
|
831
|
|
|
|
|
|
|
read FILE, $payload, $c->sizeof('packet.payload'); |
|
832
|
|
|
|
|
|
|
|
|
833
|
|
|
|
|
|
|
$packet = { |
|
834
|
|
|
|
|
|
|
header => 4711, |
|
835
|
|
|
|
|
|
|
flags => 0xf00f, |
|
836
|
|
|
|
|
|
|
payload => $payload, |
|
837
|
|
|
|
|
|
|
}; |
|
838
|
|
|
|
|
|
|
|
|
839
|
|
|
|
|
|
|
$binary = $c->pack('packet', $packet); |
|
840
|
|
|
|
|
|
|
|
|
841
|
|
|
|
|
|
|
print hexdump(data => $binary); |
|
842
|
|
|
|
|
|
|
|
|
843
|
|
|
|
|
|
|
This would print something like: |
|
844
|
|
|
|
|
|
|
|
|
845
|
|
|
|
|
|
|
0x0000 : 12 67 F0 0F 6E 6F 0A 6E 6F 0A 6E 6F 0A 6E 6F 0A : .g..no.no.no.no. |
|
846
|
|
|
|
|
|
|
0x0010 : 6E 6F 0A 6E 6F 0A 6E 6F 0A 6E 6F 0A 6E 6F 0A 6E : no.no.no.no.no.n |
|
847
|
|
|
|
|
|
|
|
|
848
|
|
|
|
|
|
|
For obvious reasons, it is not allowed to attach a C<Format> tag |
|
849
|
|
|
|
|
|
|
to bitfield members. Trying to do so will result in an exception |
|
850
|
|
|
|
|
|
|
being thrown by the L<C<tag>|/"tag"> method. |
|
851
|
|
|
|
|
|
|
|
|
852
|
|
|
|
|
|
|
=head2 The ByteOrder Tag |
|
853
|
|
|
|
|
|
|
|
|
854
|
|
|
|
|
|
|
The C<ByteOrder> tag allows you to override the byte order of |
|
855
|
|
|
|
|
|
|
certain types or members. The implementation of this tag is |
|
856
|
|
|
|
|
|
|
considered B<experimental> and may be subject to changes in the |
|
857
|
|
|
|
|
|
|
future. |
|
858
|
|
|
|
|
|
|
|
|
859
|
|
|
|
|
|
|
Usually it doesn't make much sense to override the byte order, |
|
860
|
|
|
|
|
|
|
but there may be applications where a sub-structure is packed |
|
861
|
|
|
|
|
|
|
in a different byte order than the surrounding structure. |
|
862
|
|
|
|
|
|
|
|
|
863
|
|
|
|
|
|
|
Take, for example, the following code: |
|
864
|
|
|
|
|
|
|
|
|
865
|
|
|
|
|
|
|
$c = Convert::Binary::C->new(ByteOrder => 'BigEndian', |
|
866
|
|
|
|
|
|
|
OrderMembers => 1); |
|
867
|
|
|
|
|
|
|
$c->parse(<<'ENDC'); |
|
868
|
|
|
|
|
|
|
|
|
869
|
|
|
|
|
|
|
typedef unsigned short u_16; |
|
870
|
|
|
|
|
|
|
|
|
871
|
|
|
|
|
|
|
struct coords_3d { |
|
872
|
|
|
|
|
|
|
long x, y, z; |
|
873
|
|
|
|
|
|
|
}; |
|
874
|
|
|
|
|
|
|
|
|
875
|
|
|
|
|
|
|
struct coords_msg { |
|
876
|
|
|
|
|
|
|
u_16 header; |
|
877
|
|
|
|
|
|
|
u_16 length; |
|
878
|
|
|
|
|
|
|
struct coords_3d coords; |
|
879
|
|
|
|
|
|
|
}; |
|
880
|
|
|
|
|
|
|
|
|
881
|
|
|
|
|
|
|
ENDC |
|
882
|
|
|
|
|
|
|
|
|
883
|
|
|
|
|
|
|
Assume that while C<coords_msg> is big endian, the embedded |
|
884
|
|
|
|
|
|
|
coordinates C<coords_3d> are stored in little endian format |
|
885
|
|
|
|
|
|
|
for some reason. In C, you'll have to handle this manually. |
|
886
|
|
|
|
|
|
|
|
|
887
|
|
|
|
|
|
|
But using Convert::Binary::C, you can simply attach |
|
888
|
|
|
|
|
|
|
a C<ByteOrder> tag to either the C<coords_3d> structure or to |
|
889
|
|
|
|
|
|
|
the C<coords> member of the C<coords_msg> structure. Both |
|
890
|
|
|
|
|
|
|
will work in this case. The only difference is that if you |
|
891
|
|
|
|
|
|
|
tag the C<coords> member, C<coords_3d> will only be treated |
|
892
|
|
|
|
|
|
|
as little endian if you L<C<pack>|/"pack"> or L<C<unpack>|/"unpack"> the |
|
893
|
|
|
|
|
|
|
C<coords_msg> structure. (BTW, you could also tag all members |
|
894
|
|
|
|
|
|
|
of C<coords_3d> individually, but that would be inefficient.) |
|
895
|
|
|
|
|
|
|
|
|
896
|
|
|
|
|
|
|
So, let's attach the C<ByteOrder> tag to the C<coords> member: |
|
897
|
|
|
|
|
|
|
|
|
898
|
|
|
|
|
|
|
$c->tag('coords_msg.coords', ByteOrder => 'LittleEndian'); |
|
899
|
|
|
|
|
|
|
|
|
900
|
|
|
|
|
|
|
Assume the following binary message: |
|
901
|
|
|
|
|
|
|
|
|
902
|
|
|
|
|
|
|
0x0000 : 00 2A 00 0C FF FF FF FF 02 00 00 00 2A 00 00 00 : .*..........*... |
|
903
|
|
|
|
|
|
|
|
|
904
|
|
|
|
|
|
|
If you unpack this message... |
|
905
|
|
|
|
|
|
|
|
|
906
|
|
|
|
|
|
|
$msg = $c->unpack('coords_msg', $binary); |
|
907
|
|
|
|
|
|
|
|
|
908
|
|
|
|
|
|
|
...you will get the following data structure: |
|
909
|
|
|
|
|
|
|
|
|
910
|
|
|
|
|
|
|
$msg = { |
|
911
|
|
|
|
|
|
|
'header' => 42, |
|
912
|
|
|
|
|
|
|
'length' => 12, |
|
913
|
|
|
|
|
|
|
'coords' => { |
|
914
|
|
|
|
|
|
|
'x' => -1, |
|
915
|
|
|
|
|
|
|
'y' => 2, |
|
916
|
|
|
|
|
|
|
'z' => 42 |
|
917
|
|
|
|
|
|
|
} |
|
918
|
|
|
|
|
|
|
}; |
|
919
|
|
|
|
|
|
|
|
|
920
|
|
|
|
|
|
|
Without the C<ByteOrder> tag, you would get: |
|
921
|
|
|
|
|
|
|
|
|
922
|
|
|
|
|
|
|
$msg = { |
|
923
|
|
|
|
|
|
|
'header' => 42, |
|
924
|
|
|
|
|
|
|
'length' => 12, |
|
925
|
|
|
|
|
|
|
'coords' => { |
|
926
|
|
|
|
|
|
|
'x' => -1, |
|
927
|
|
|
|
|
|
|
'y' => 33554432, |
|
928
|
|
|
|
|
|
|
'z' => 704643072 |
|
929
|
|
|
|
|
|
|
} |
|
930
|
|
|
|
|
|
|
}; |
|
931
|
|
|
|
|
|
|
|
|
932
|
|
|
|
|
|
|
The C<ByteOrder> tag is a I<recursive> tag, i.e. it applies |
|
933
|
|
|
|
|
|
|
to all children of the tagged object recursively. Of course, |
|
934
|
|
|
|
|
|
|
it is also possible to override a C<ByteOrder> tag by attaching |
|
935
|
|
|
|
|
|
|
another C<ByteOrder> tag to a child type. Confused? Here's an |
|
936
|
|
|
|
|
|
|
example. In addition to tagging the C<coords> member as little |
|
937
|
|
|
|
|
|
|
endian, we now tag C<coords_3d.y> as big endian: |
|
938
|
|
|
|
|
|
|
|
|
939
|
|
|
|
|
|
|
$c->tag('coords_3d.y', ByteOrder => 'BigEndian'); |
|
940
|
|
|
|
|
|
|
$msg = $c->unpack('coords_msg', $binary); |
|
941
|
|
|
|
|
|
|
|
|
942
|
|
|
|
|
|
|
This will return the following data structure: |
|
943
|
|
|
|
|
|
|
|
|
944
|
|
|
|
|
|
|
$msg = { |
|
945
|
|
|
|
|
|
|
'header' => 42, |
|
946
|
|
|
|
|
|
|
'length' => 12, |
|
947
|
|
|
|
|
|
|
'coords' => { |
|
948
|
|
|
|
|
|
|
'x' => -1, |
|
949
|
|
|
|
|
|
|
'y' => 33554432, |
|
950
|
|
|
|
|
|
|
'z' => 42 |
|
951
|
|
|
|
|
|
|
} |
|
952
|
|
|
|
|
|
|
}; |
|
953
|
|
|
|
|
|
|
|
|
954
|
|
|
|
|
|
|
Note that if you tag both a type and a member of that type |
|
955
|
|
|
|
|
|
|
within a compound, the tag attached to the type itself has |
|
956
|
|
|
|
|
|
|
higher precedence. Using the example above, if you would attach |
|
957
|
|
|
|
|
|
|
a C<ByteOrder> tag to both C<coords_msg.coords> and C<coords_3d>, |
|
958
|
|
|
|
|
|
|
the tag attached to C<coords_3d> would always win. |
|
959
|
|
|
|
|
|
|
|
|
960
|
|
|
|
|
|
|
Also note that the C<ByteOrder> tag might not work as expected |
|
961
|
|
|
|
|
|
|
along with bitfields, which is why the implementation is considered |
|
962
|
|
|
|
|
|
|
experimental. Bitfields are currently B<not> affected by |
|
963
|
|
|
|
|
|
|
the C<ByteOrder> tag at all. This is because the byte order |
|
964
|
|
|
|
|
|
|
would affect the bitfield layout, and a consistent implementation |
|
965
|
|
|
|
|
|
|
supporting multiple layouts of the same struct would be quite |
|
966
|
|
|
|
|
|
|
bulky and probably slow down the whole module. |
|
967
|
|
|
|
|
|
|
|
|
968
|
|
|
|
|
|
|
If you really need the correct behaviour, you can use the |
|
969
|
|
|
|
|
|
|
following trick: |
|
970
|
|
|
|
|
|
|
|
|
971
|
|
|
|
|
|
|
$le = Convert::Binary::C->new(ByteOrder => 'LittleEndian'); |
|
972
|
|
|
|
|
|
|
|
|
973
|
|
|
|
|
|
|
$le->parse(<<'ENDC'); |
|
974
|
|
|
|
|
|
|
|
|
975
|
|
|
|
|
|
|
typedef unsigned short u_16; |
|
976
|
|
|
|
|
|
|
typedef unsigned long u_32; |
|
977
|
|
|
|
|
|
|
|
|
978
|
|
|
|
|
|
|
struct message { |
|
979
|
|
|
|
|
|
|
u_16 header; |
|
980
|
|
|
|
|
|
|
u_16 length; |
|
981
|
|
|
|
|
|
|
struct { |
|
982
|
|
|
|
|
|
|
u_32 a; |
|
983
|
|
|
|
|
|
|
u_32 b; |
|
984
|
|
|
|
|
|
|
u_32 c : 7; |
|
985
|
|
|
|
|
|
|
u_32 d : 5; |
|
986
|
|
|
|
|
|
|
u_32 e : 20; |
|
987
|
|
|
|
|
|
|
} data; |
|
988
|
|
|
|
|
|
|
}; |
|
989
|
|
|
|
|
|
|
|
|
990
|
|
|
|
|
|
|
ENDC |
|
991
|
|
|
|
|
|
|
|
|
992
|
|
|
|
|
|
|
$be = $le->clone->ByteOrder('BigEndian'); |
|
993
|
|
|
|
|
|
|
|
|
994
|
|
|
|
|
|
|
$le->tag('message.data', Format => 'Binary', Hooks => { |
|
995
|
|
|
|
|
|
|
unpack => sub { $be->unpack('message.data', @_) }, |
|
996
|
|
|
|
|
|
|
pack => sub { $be->pack('message.data', @_) }, |
|
997
|
|
|
|
|
|
|
}); |
|
998
|
|
|
|
|
|
|
|
|
999
|
|
|
|
|
|
|
|
|
1000
|
|
|
|
|
|
|
$msg = $le->unpack('message', $binary); |
|
1001
|
|
|
|
|
|
|
|
|
1002
|
|
|
|
|
|
|
This uses the L<C<Format>|/"The Format Tag"> and L<C<Hooks>|/"The Hooks Tag"> tags |
|
1003
|
|
|
|
|
|
|
along with a big endian L<C<clone>|/"clone"> of the original |
|
1004
|
|
|
|
|
|
|
little endian object. It attaches hooks to the little endian |
|
1005
|
|
|
|
|
|
|
object and in the hooks it uses the big endian object |
|
1006
|
|
|
|
|
|
|
to L<C<pack>|/"pack"> and L<C<unpack>|/"unpack"> the binary data. |
|
1007
|
|
|
|
|
|
|
|
|
1008
|
|
|
|
|
|
|
=head2 The Dimension Tag |
|
1009
|
|
|
|
|
|
|
|
|
1010
|
|
|
|
|
|
|
The C<Dimension> tag allows you to override the declared dimension |
|
1011
|
|
|
|
|
|
|
of an array for packing or unpacking data. The implementation of |
|
1012
|
|
|
|
|
|
|
this tag is considered B<very experimental> and will B<definitely change> in |
|
1013
|
|
|
|
|
|
|
a future release. |
|
1014
|
|
|
|
|
|
|
|
|
1015
|
|
|
|
|
|
|
That being said, the C<Dimension> tag is primarily useful to support |
|
1016
|
|
|
|
|
|
|
variable length arrays. Usually, you have to write the following code |
|
1017
|
|
|
|
|
|
|
for such a variable length array in C: |
|
1018
|
|
|
|
|
|
|
|
|
1019
|
|
|
|
|
|
|
struct c_message |
|
1020
|
|
|
|
|
|
|
{ |
|
1021
|
|
|
|
|
|
|
unsigned count; |
|
1022
|
|
|
|
|
|
|
char data[1]; |
|
1023
|
|
|
|
|
|
|
}; |
|
1024
|
|
|
|
|
|
|
|
|
1025
|
|
|
|
|
|
|
So, because you cannot declare an empty array, you declare an array |
|
1026
|
|
|
|
|
|
|
with a single element. If you have a ISO-C99 compliant compiler, |
|
1027
|
|
|
|
|
|
|
you can write this code instead: |
|
1028
|
|
|
|
|
|
|
|
|
1029
|
|
|
|
|
|
|
struct c99_message |
|
1030
|
|
|
|
|
|
|
{ |
|
1031
|
|
|
|
|
|
|
unsigned count; |
|
1032
|
|
|
|
|
|
|
char data[]; |
|
1033
|
|
|
|
|
|
|
}; |
|
1034
|
|
|
|
|
|
|
|
|
1035
|
|
|
|
|
|
|
This explicitly tells the compiler that C<data> is a flexible array |
|
1036
|
|
|
|
|
|
|
member. Convert::Binary::C already uses this information to |
|
1037
|
|
|
|
|
|
|
handle L<flexible array members|/"FLEXIBLE ARRAY MEMBERS AND INCOMPLETE TYPES"> in |
|
1038
|
|
|
|
|
|
|
a special way. |
|
1039
|
|
|
|
|
|
|
|
|
1040
|
|
|
|
|
|
|
As you can see in the following example, the two types are treated |
|
1041
|
|
|
|
|
|
|
differently: |
|
1042
|
|
|
|
|
|
|
|
|
1043
|
|
|
|
|
|
|
$data = pack 'NC*', 3, 1..8; |
|
1044
|
|
|
|
|
|
|
$uc = $c->unpack('c_message', $data); |
|
1045
|
|
|
|
|
|
|
$uc99 = $c->unpack('c99_message', $data); |
|
1046
|
|
|
|
|
|
|
|
|
1047
|
|
|
|
|
|
|
This will result in: |
|
1048
|
|
|
|
|
|
|
|
|
1049
|
|
|
|
|
|
|
$uc = {'count' => 3,'data' => [1]}; |
|
1050
|
|
|
|
|
|
|
$uc99 = {'count' => 3,'data' => [1,2,3,4,5,6,7,8]}; |
|
1051
|
|
|
|
|
|
|
|
|
1052
|
|
|
|
|
|
|
However, only few compilers support ISO-C99, and you probably don't want |
|
1053
|
|
|
|
|
|
|
to change your existing code only to get some extra features when |
|
1054
|
|
|
|
|
|
|
using Convert::Binary::C. |
|
1055
|
|
|
|
|
|
|
|
|
1056
|
|
|
|
|
|
|
So it is possible to attach a tag to the C<data> member of |
|
1057
|
|
|
|
|
|
|
the C<c_message> struct that tells Convert::Binary::C to treat |
|
1058
|
|
|
|
|
|
|
the array as if it were flexible: |
|
1059
|
|
|
|
|
|
|
|
|
1060
|
|
|
|
|
|
|
$c->tag('c_message.data', Dimension => '*'); |
|
1061
|
|
|
|
|
|
|
|
|
1062
|
|
|
|
|
|
|
Now both C<c_message> and C<c99_message> will behave exactly the |
|
1063
|
|
|
|
|
|
|
same when using L<C<pack>|/"pack"> or L<C<unpack>|/"unpack">. |
|
1064
|
|
|
|
|
|
|
Repeating the above code: |
|
1065
|
|
|
|
|
|
|
|
|
1066
|
|
|
|
|
|
|
$uc = $c->unpack('c_message', $data); |
|
1067
|
|
|
|
|
|
|
|
|
1068
|
|
|
|
|
|
|
This will result in: |
|
1069
|
|
|
|
|
|
|
|
|
1070
|
|
|
|
|
|
|
$uc = {'count' => 3,'data' => [1,2,3,4,5,6,7,8]}; |
|
1071
|
|
|
|
|
|
|
|
|
1072
|
|
|
|
|
|
|
But there's more you can do. Even though it probably doesn't |
|
1073
|
|
|
|
|
|
|
make much sense, you can tag a fixed dimension to an array: |
|
1074
|
|
|
|
|
|
|
|
|
1075
|
|
|
|
|
|
|
$c->tag('c_message.data', Dimension => '5'); |
|
1076
|
|
|
|
|
|
|
|
|
1077
|
|
|
|
|
|
|
This will obviously result in: |
|
1078
|
|
|
|
|
|
|
|
|
1079
|
|
|
|
|
|
|
$uc = {'count' => 3,'data' => [1,2,3,4,5]}; |
|
1080
|
|
|
|
|
|
|
|
|
1081
|
|
|
|
|
|
|
A more useful way to use the C<Dimension> tag is to set it to |
|
1082
|
|
|
|
|
|
|
the name of a member in the same compound: |
|
1083
|
|
|
|
|
|
|
|
|
1084
|
|
|
|
|
|
|
$c->tag('c_message.data', Dimension => 'count'); |
|
1085
|
|
|
|
|
|
|
|
|
1086
|
|
|
|
|
|
|
Convert::Binary::C will now use the value of that member to |
|
1087
|
|
|
|
|
|
|
determine the size of the array, so unpacking will result in: |
|
1088
|
|
|
|
|
|
|
|
|
1089
|
|
|
|
|
|
|
$uc = {'count' => 3,'data' => [1,2,3]}; |
|
1090
|
|
|
|
|
|
|
|
|
1091
|
|
|
|
|
|
|
Of course, you can also tag flexible array members. And yes, |
|
1092
|
|
|
|
|
|
|
it's also possible to use more complex member expressions: |
|
1093
|
|
|
|
|
|
|
|
|
1094
|
|
|
|
|
|
|
$c->parse(<<ENDC); |
|
1095
|
|
|
|
|
|
|
struct msg_header |
|
1096
|
|
|
|
|
|
|
{ |
|
1097
|
|
|
|
|
|
|
unsigned len[2]; |
|
1098
|
|
|
|
|
|
|
}; |
|
1099
|
|
|
|
|
|
|
|
|
1100
|
|
|
|
|
|
|
struct more_complex |
|
1101
|
|
|
|
|
|
|
{ |
|
1102
|
|
|
|
|
|
|
struct msg_header hdr; |
|
1103
|
|
|
|
|
|
|
char data[]; |
|
1104
|
|
|
|
|
|
|
}; |
|
1105
|
|
|
|
|
|
|
ENDC |
|
1106
|
|
|
|
|
|
|
|
|
1107
|
|
|
|
|
|
|
$data = pack 'NNC*', 42, 7, 1 .. 10; |
|
1108
|
|
|
|
|
|
|
|
|
1109
|
|
|
|
|
|
|
$c->tag('more_complex.data', Dimension => 'hdr.len[1]'); |
|
1110
|
|
|
|
|
|
|
|
|
1111
|
|
|
|
|
|
|
$u = $c->unpack('more_complex', $data); |
|
1112
|
|
|
|
|
|
|
|
|
1113
|
|
|
|
|
|
|
The result will be: |
|
1114
|
|
|
|
|
|
|
|
|
1115
|
|
|
|
|
|
|
$u = { |
|
1116
|
|
|
|
|
|
|
'hdr' => { |
|
1117
|
|
|
|
|
|
|
'len' => [ |
|
1118
|
|
|
|
|
|
|
42, |
|
1119
|
|
|
|
|
|
|
7 |
|
1120
|
|
|
|
|
|
|
] |
|
1121
|
|
|
|
|
|
|
}, |
|
1122
|
|
|
|
|
|
|
'data' => [ |
|
1123
|
|
|
|
|
|
|
1, |
|
1124
|
|
|
|
|
|
|
2, |
|
1125
|
|
|
|
|
|
|
3, |
|
1126
|
|
|
|
|
|
|
4, |
|
1127
|
|
|
|
|
|
|
5, |
|
1128
|
|
|
|
|
|
|
6, |
|
1129
|
|
|
|
|
|
|
7 |
|
1130
|
|
|
|
|
|
|
] |
|
1131
|
|
|
|
|
|
|
}; |
|
1132
|
|
|
|
|
|
|
|
|
1133
|
|
|
|
|
|
|
By the way, it's also possible to tag arrays that are not |
|
1134
|
|
|
|
|
|
|
embedded inside a compound: |
|
1135
|
|
|
|
|
|
|
|
|
1136
|
|
|
|
|
|
|
$c->parse(<<ENDC); |
|
1137
|
|
|
|
|
|
|
typedef unsigned short short_array[]; |
|
1138
|
|
|
|
|
|
|
ENDC |
|
1139
|
|
|
|
|
|
|
|
|
1140
|
|
|
|
|
|
|
$c->tag('short_array', Dimension => '5'); |
|
1141
|
|
|
|
|
|
|
|
|
1142
|
|
|
|
|
|
|
$u = $c->unpack('short_array', $data); |
|
1143
|
|
|
|
|
|
|
|
|
1144
|
|
|
|
|
|
|
Resulting in: |
|
1145
|
|
|
|
|
|
|
|
|
1146
|
|
|
|
|
|
|
$u = [0,42,0,7,258]; |
|
1147
|
|
|
|
|
|
|
|
|
1148
|
|
|
|
|
|
|
The final and most powerful way to define a C<Dimension> tag is |
|
1149
|
|
|
|
|
|
|
to pass it a subroutine reference. The referenced subroutine can |
|
1150
|
|
|
|
|
|
|
execute whatever code is neccessary to determine the size of the |
|
1151
|
|
|
|
|
|
|
tagged array: |
|
1152
|
|
|
|
|
|
|
|
|
1153
|
|
|
|
|
|
|
sub get_size |
|
1154
|
|
|
|
|
|
|
{ |
|
1155
|
|
|
|
|
|
|
my $m = shift; |
|
1156
|
|
|
|
|
|
|
return $m->{hdr}{len}[0] / $m->{hdr}{len}[1]; |
|
1157
|
|
|
|
|
|
|
} |
|
1158
|
|
|
|
|
|
|
|
|
1159
|
|
|
|
|
|
|
$c->tag('more_complex.data', Dimension => \&get_size); |
|
1160
|
|
|
|
|
|
|
|
|
1161
|
|
|
|
|
|
|
$u = $c->unpack('more_complex', $data); |
|
1162
|
|
|
|
|
|
|
|
|
1163
|
|
|
|
|
|
|
As you can guess from the above code, the subroutine is being passed |
|
1164
|
|
|
|
|
|
|
a reference to hash that stores the already unpacked part of the |
|
1165
|
|
|
|
|
|
|
compound embedding the tagged array. This is the result: |
|
1166
|
|
|
|
|
|
|
|
|
1167
|
|
|
|
|
|
|
$u = { |
|
1168
|
|
|
|
|
|
|
'hdr' => { |
|
1169
|
|
|
|
|
|
|
'len' => [ |
|
1170
|
|
|
|
|
|
|
42, |
|
1171
|
|
|
|
|
|
|
7 |
|
1172
|
|
|
|
|
|
|
] |
|
1173
|
|
|
|
|
|
|
}, |
|
1174
|
|
|
|
|
|
|
'data' => [ |
|
1175
|
|
|
|
|
|
|
1, |
|
1176
|
|
|
|
|
|
|
2, |
|
1177
|
|
|
|
|
|
|
3, |
|
1178
|
|
|
|
|
|
|
4, |
|
1179
|
|
|
|
|
|
|
5, |
|
1180
|
|
|
|
|
|
|
6 |
|
1181
|
|
|
|
|
|
|
] |
|
1182
|
|
|
|
|
|
|
}; |
|
1183
|
|
|
|
|
|
|
|
|
1184
|
|
|
|
|
|
|
You can also pass custom arguments to the subroutines by using |
|
1185
|
|
|
|
|
|
|
the L<C<arg>|/"arg"> method. This is similar to the functionality |
|
1186
|
|
|
|
|
|
|
offered by the L<C<Hooks>|/"The Hooks Tag"> tag. |
|
1187
|
|
|
|
|
|
|
|
|
1188
|
|
|
|
|
|
|
Of course, all that also works for the L<C<pack>|/"pack"> method |
|
1189
|
|
|
|
|
|
|
as well. |
|
1190
|
|
|
|
|
|
|
|
|
1191
|
|
|
|
|
|
|
However, the current implementation has at least one shortcomings, |
|
1192
|
|
|
|
|
|
|
which is why it's experimental: The C<Dimension> tag doesn't impact |
|
1193
|
|
|
|
|
|
|
compound layout. This means that while you can alter the size of an |
|
1194
|
|
|
|
|
|
|
array in the middle of a compound, the offset of the members after |
|
1195
|
|
|
|
|
|
|
that array won't be impacted. I'd rather like to see the layout adapt |
|
1196
|
|
|
|
|
|
|
dynamically, so this is what I'm hoping to implement in the future. |
|
1197
|
|
|
|
|
|
|
|
|
1198
|
|
|
|
|
|
|
=head2 The Hooks Tag |
|
1199
|
|
|
|
|
|
|
|
|
1200
|
|
|
|
|
|
|
Hooks are a special kind of tag that can be extremely useful. |
|
1201
|
|
|
|
|
|
|
|
|
1202
|
|
|
|
|
|
|
Using hooks, you can easily override the |
|
1203
|
|
|
|
|
|
|
way L<C<pack>|/"pack"> and L<C<unpack>|/"unpack"> handle data |
|
1204
|
|
|
|
|
|
|
using your own subroutines. |
|
1205
|
|
|
|
|
|
|
If you define hooks for a certain data type, each time this |
|
1206
|
|
|
|
|
|
|
data type is processed the corresponding hook will be called |
|
1207
|
|
|
|
|
|
|
to allow you to modify that data. |
|
1208
|
|
|
|
|
|
|
|
|
1209
|
|
|
|
|
|
|
=head3 Basic Hooks |
|
1210
|
|
|
|
|
|
|
|
|
1211
|
|
|
|
|
|
|
Here's an example. Let's assume the following C code has been |
|
1212
|
|
|
|
|
|
|
parsed: |
|
1213
|
|
|
|
|
|
|
|
|
1214
|
|
|
|
|
|
|
typedef unsigned long u_32; |
|
1215
|
|
|
|
|
|
|
typedef u_32 ProtoId; |
|
1216
|
|
|
|
|
|
|
typedef ProtoId MyProtoId; |
|
1217
|
|
|
|
|
|
|
|
|
1218
|
|
|
|
|
|
|
struct MsgHeader { |
|
1219
|
|
|
|
|
|
|
MyProtoId id; |
|
1220
|
|
|
|
|
|
|
u_32 len; |
|
1221
|
|
|
|
|
|
|
}; |
|
1222
|
|
|
|
|
|
|
|
|
1223
|
|
|
|
|
|
|
struct String { |
|
1224
|
|
|
|
|
|
|
u_32 len; |
|
1225
|
|
|
|
|
|
|
char buf[]; |
|
1226
|
|
|
|
|
|
|
}; |
|
1227
|
|
|
|
|
|
|
|
|
1228
|
|
|
|
|
|
|
You could now use the types above and, for example, unpack |
|
1229
|
|
|
|
|
|
|
binary data representing a C<MsgHeader> like this: |
|
1230
|
|
|
|
|
|
|
|
|
1231
|
|
|
|
|
|
|
$msg_header = $c->unpack('MsgHeader', $data); |
|
1232
|
|
|
|
|
|
|
|
|
1233
|
|
|
|
|
|
|
This would give you: |
|
1234
|
|
|
|
|
|
|
|
|
1235
|
|
|
|
|
|
|
$msg_header = { |
|
1236
|
|
|
|
|
|
|
'len' => 13, |
|
1237
|
|
|
|
|
|
|
'id' => 42 |
|
1238
|
|
|
|
|
|
|
}; |
|
1239
|
|
|
|
|
|
|
|
|
1240
|
|
|
|
|
|
|
Instead of dealing with C<ProtoId>'s as integers, you would |
|
1241
|
|
|
|
|
|
|
rather like to have them as clear text. You could provide |
|
1242
|
|
|
|
|
|
|
subroutines to convert between clear text and integers: |
|
1243
|
|
|
|
|
|
|
|
|
1244
|
|
|
|
|
|
|
%proto = ( |
|
1245
|
|
|
|
|
|
|
CATS => 1, |
|
1246
|
|
|
|
|
|
|
DOGS => 42, |
|
1247
|
|
|
|
|
|
|
HEDGEHOGS => 4711, |
|
1248
|
|
|
|
|
|
|
); |
|
1249
|
|
|
|
|
|
|
|
|
1250
|
|
|
|
|
|
|
%rproto = reverse %proto; |
|
1251
|
|
|
|
|
|
|
|
|
1252
|
|
|
|
|
|
|
sub ProtoId_unpack { |
|
1253
|
|
|
|
|
|
|
$rproto{$_[0]} || 'unknown protocol' |
|
1254
|
|
|
|
|
|
|
} |
|
1255
|
|
|
|
|
|
|
|
|
1256
|
|
|
|
|
|
|
sub ProtoId_pack { |
|
1257
|
|
|
|
|
|
|
$proto{$_[0]} or die 'unknown protocol' |
|
1258
|
|
|
|
|
|
|
} |
|
1259
|
|
|
|
|
|
|
|
|
1260
|
|
|
|
|
|
|
You can now register these subroutines by attaching a C<Hooks> tag |
|
1261
|
|
|
|
|
|
|
to C<ProtoId> using the L<C<tag>|/"tag"> method: |
|
1262
|
|
|
|
|
|
|
|
|
1263
|
|
|
|
|
|
|
$c->tag('ProtoId', Hooks => { pack => \&ProtoId_pack, |
|
1264
|
|
|
|
|
|
|
unpack => \&ProtoId_unpack }); |
|
1265
|
|
|
|
|
|
|
|
|
1266
|
|
|
|
|
|
|
Doing exactly the same unpack on C<MsgHeader> again would |
|
1267
|
|
|
|
|
|
|
now return: |
|
1268
|
|
|
|
|
|
|
|
|
1269
|
|
|
|
|
|
|
$msg_header = { |
|
1270
|
|
|
|
|
|
|
'len' => 13, |
|
1271
|
|
|
|
|
|
|
'id' => 'DOGS' |
|
1272
|
|
|
|
|
|
|
}; |
|
1273
|
|
|
|
|
|
|
|
|
1274
|
|
|
|
|
|
|
Actually, if you don't need the reverse operation, you don't even |
|
1275
|
|
|
|
|
|
|
have to register a C<pack> hook. Or, even better, you can have a |
|
1276
|
|
|
|
|
|
|
more intelligent C<unpack> hook that creates a dual-typed variable: |
|
1277
|
|
|
|
|
|
|
|
|
1278
|
|
|
|
|
|
|
use Scalar::Util qw(dualvar); |
|
1279
|
|
|
|
|
|
|
|
|
1280
|
|
|
|
|
|
|
sub ProtoId_unpack2 { |
|
1281
|
|
|
|
|
|
|
dualvar $_[0], $rproto{$_[0]} || 'unknown protocol' |
|
1282
|
|
|
|
|
|
|
} |
|
1283
|
|
|
|
|
|
|
|
|
1284
|
|
|
|
|
|
|
$c->tag('ProtoId', Hooks => { unpack => \&ProtoId_unpack2 }); |
|
1285
|
|
|
|
|
|
|
|
|
1286
|
|
|
|
|
|
|
$msg_header = $c->unpack('MsgHeader', $data); |
|
1287
|
|
|
|
|
|
|
|
|
1288
|
|
|
|
|
|
|
Just as before, this would print |
|
1289
|
|
|
|
|
|
|
|
|
1290
|
|
|
|
|
|
|
$msg_header = { |
|
1291
|
|
|
|
|
|
|
'len' => 13, |
|
1292
|
|
|
|
|
|
|
'id' => 'DOGS' |
|
1293
|
|
|
|
|
|
|
}; |
|
1294
|
|
|
|
|
|
|
|
|
1295
|
|
|
|
|
|
|
but without requiring a C<pack> hook for packing, at least as |
|
1296
|
|
|
|
|
|
|
long as you keep the variable dual-typed. |
|
1297
|
|
|
|
|
|
|
|
|
1298
|
|
|
|
|
|
|
Hooks are usually called with exactly one argument, which is the |
|
1299
|
|
|
|
|
|
|
data that should be processed (see L<"Advanced Hooks"> for details |
|
1300
|
|
|
|
|
|
|
on how to customize hook arguments). They are called in scalar |
|
1301
|
|
|
|
|
|
|
context and expected to return the processed data. |
|
1302
|
|
|
|
|
|
|
|
|
1303
|
|
|
|
|
|
|
To get rid of registered hooks, you can either undefine only |
|
1304
|
|
|
|
|
|
|
certain hooks |
|
1305
|
|
|
|
|
|
|
|
|
1306
|
|
|
|
|
|
|
$c->tag('ProtoId', Hooks => { pack => undef }); |
|
1307
|
|
|
|
|
|
|
|
|
1308
|
|
|
|
|
|
|
or all hooks: |
|
1309
|
|
|
|
|
|
|
|
|
1310
|
|
|
|
|
|
|
$c->tag('ProtoId', Hooks => undef); |
|
1311
|
|
|
|
|
|
|
|
|
1312
|
|
|
|
|
|
|
Of course, hooks are not restricted to handling integer values. |
|
1313
|
|
|
|
|
|
|
You could just as well attach hooks for the C<String> struct from |
|
1314
|
|
|
|
|
|
|
the code above. A useful example would be to have these hooks: |
|
1315
|
|
|
|
|
|
|
|
|
1316
|
|
|
|
|
|
|
sub string_unpack { |
|
1317
|
|
|
|
|
|
|
my $s = shift; |
|
1318
|
|
|
|
|
|
|
pack "c$s->{len}", @{$s->{buf}}; |
|
1319
|
|
|
|
|
|
|
} |
|
1320
|
|
|
|
|
|
|
|
|
1321
|
|
|
|
|
|
|
sub string_pack { |
|
1322
|
|
|
|
|
|
|
my $s = shift; |
|
1323
|
|
|
|
|
|
|
return { |
|
1324
|
|
|
|
|
|
|
len => length $s, |
|
1325
|
|
|
|
|
|
|
buf => [ unpack 'c*', $s ], |
|
1326
|
|
|
|
|
|
|
} |
|
1327
|
|
|
|
|
|
|
} |
|
1328
|
|
|
|
|
|
|
|
|
1329
|
|
|
|
|
|
|
(Don't be confused by the fact that the C<unpack> hook |
|
1330
|
|
|
|
|
|
|
uses C<pack> and the C<pack> hook uses C<unpack>. |
|
1331
|
|
|
|
|
|
|
And also see L<"Advanced Hooks"> for a more clever approach.) |
|
1332
|
|
|
|
|
|
|
|
|
1333
|
|
|
|
|
|
|
While you would normally get the following output when unpacking |
|
1334
|
|
|
|
|
|
|
a C<String> |
|
1335
|
|
|
|
|
|
|
|
|
1336
|
|
|
|
|
|
|
$string = { |
|
1337
|
|
|
|
|
|
|
'len' => 12, |
|
1338
|
|
|
|
|
|
|
'buf' => [ |
|
1339
|
|
|
|
|
|
|
72, |
|
1340
|
|
|
|
|
|
|
101, |
|
1341
|
|
|
|
|
|
|
108, |
|
1342
|
|
|
|
|
|
|
108, |
|
1343
|
|
|
|
|
|
|
111, |
|
1344
|
|
|
|
|
|
|
32, |
|
1345
|
|
|
|
|
|
|
87, |
|
1346
|
|
|
|
|
|
|
111, |
|
1347
|
|
|
|
|
|
|
114, |
|
1348
|
|
|
|
|
|
|
108, |
|
1349
|
|
|
|
|
|
|
100, |
|
1350
|
|
|
|
|
|
|
33 |
|
1351
|
|
|
|
|
|
|
] |
|
1352
|
|
|
|
|
|
|
}; |
|
1353
|
|
|
|
|
|
|
|
|
1354
|
|
|
|
|
|
|
you could just register the hooks using |
|
1355
|
|
|
|
|
|
|
|
|
1356
|
|
|
|
|
|
|
$c->tag('String', Hooks => { pack => \&string_pack, |
|
1357
|
|
|
|
|
|
|
unpack => \&string_unpack }); |
|
1358
|
|
|
|
|
|
|
|
|
1359
|
|
|
|
|
|
|
and you would get a nice human-readable Perl string: |
|
1360
|
|
|
|
|
|
|
|
|
1361
|
|
|
|
|
|
|
$string = 'Hello World!'; |
|
1362
|
|
|
|
|
|
|
|
|
1363
|
|
|
|
|
|
|
Packing a string turns out to be just as easy: |
|
1364
|
|
|
|
|
|
|
|
|
1365
|
|
|
|
|
|
|
use Data::Hexdumper; |
|
1366
|
|
|
|
|
|
|
|
|
1367
|
|
|
|
|
|
|
$data = $c->pack('String', 'Just another Perl hacker,'); |
|
1368
|
|
|
|
|
|
|
|
|
1369
|
|
|
|
|
|
|
print hexdump(data => $data); |
|
1370
|
|
|
|
|
|
|
|
|
1371
|
|
|
|
|
|
|
This would print: |
|
1372
|
|
|
|
|
|
|
|
|
1373
|
|
|
|
|
|
|
0x0000 : 00 00 00 19 4A 75 73 74 20 61 6E 6F 74 68 65 72 : ....Just.another |
|
1374
|
|
|
|
|
|
|
0x0010 : 20 50 65 72 6C 20 68 61 63 6B 65 72 2C : .Perl.hacker, |
|
1375
|
|
|
|
|
|
|
|
|
1376
|
|
|
|
|
|
|
If you want to find out if or which hooks are registered for |
|
1377
|
|
|
|
|
|
|
a certain type, you can also use the L<C<tag>|/"tag"> method: |
|
1378
|
|
|
|
|
|
|
|
|
1379
|
|
|
|
|
|
|
$hooks = $c->tag('String', 'Hooks'); |
|
1380
|
|
|
|
|
|
|
|
|
1381
|
|
|
|
|
|
|
This would return: |
|
1382
|
|
|
|
|
|
|
|
|
1383
|
|
|
|
|
|
|
$hooks = { |
|
1384
|
|
|
|
|
|
|
'unpack' => \&string_unpack, |
|
1385
|
|
|
|
|
|
|
'pack' => \&string_pack |
|
1386
|
|
|
|
|
|
|
}; |
|
1387
|
|
|
|
|
|
|
|
|
1388
|
|
|
|
|
|
|
=head3 Advanced Hooks |
|
1389
|
|
|
|
|
|
|
|
|
1390
|
|
|
|
|
|
|
It is also possible to combine hooks with using the C<Format> tag. |
|
1391
|
|
|
|
|
|
|
This can be useful if you know better than Convert::Binary::C how |
|
1392
|
|
|
|
|
|
|
to interpret the binary data. In the previous section, we've handled |
|
1393
|
|
|
|
|
|
|
this type |
|
1394
|
|
|
|
|
|
|
|
|
1395
|
|
|
|
|
|
|
struct String { |
|
1396
|
|
|
|
|
|
|
u_32 len; |
|
1397
|
|
|
|
|
|
|
char buf[]; |
|
1398
|
|
|
|
|
|
|
}; |
|
1399
|
|
|
|
|
|
|
|
|
1400
|
|
|
|
|
|
|
with the following hooks: |
|
1401
|
|
|
|
|
|
|
|
|
1402
|
|
|
|
|
|
|
sub string_unpack { |
|
1403
|
|
|
|
|
|
|
my $s = shift; |
|
1404
|
|
|
|
|
|
|
pack "c$s->{len}", @{$s->{buf}}; |
|
1405
|
|
|
|
|
|
|
} |
|
1406
|
|
|
|
|
|
|
|
|
1407
|
|
|
|
|
|
|
sub string_pack { |
|
1408
|
|
|
|
|
|
|
my $s = shift; |
|
1409
|
|
|
|
|
|
|
return { |
|
1410
|
|
|
|
|
|
|
len => length $s, |
|
1411
|
|
|
|
|
|
|
buf => [ unpack 'c*', $s ], |
|
1412
|
|
|
|
|
|
|
} |
|
1413
|
|
|
|
|
|
|
} |
|
1414
|
|
|
|
|
|
|
|
|
1415
|
|
|
|
|
|
|
$c->tag('String', Hooks => { pack => \&string_pack, |
|
1416
|
|
|
|
|
|
|
unpack => \&string_unpack }); |
|
1417
|
|
|
|
|
|
|
|
|
1418
|
|
|
|
|
|
|
As you can see in the hook code, C<buf> is expected to be an array |
|
1419
|
|
|
|
|
|
|
of characters. For the L<C<unpack>|/"unpack"> case Convert::Binary::C |
|
1420
|
|
|
|
|
|
|
first turns the binary data into a Perl array, and then the hook packs |
|
1421
|
|
|
|
|
|
|
it back into a string. The intermediate array creation and destruction |
|
1422
|
|
|
|
|
|
|
is completely useless. |
|
1423
|
|
|
|
|
|
|
Same thing, of course, for the L<C<pack>|/"pack"> case. |
|
1424
|
|
|
|
|
|
|
|
|
1425
|
|
|
|
|
|
|
Here's a clever way to handle this. Just tag C<buf> as binary |
|
1426
|
|
|
|
|
|
|
|
|
1427
|
|
|
|
|
|
|
$c->tag('String.buf', Format => 'Binary'); |
|
1428
|
|
|
|
|
|
|
|
|
1429
|
|
|
|
|
|
|
and use the following hooks instead: |
|
1430
|
|
|
|
|
|
|
|
|
1431
|
|
|
|
|
|
|
sub string_unpack2 { |
|
1432
|
|
|
|
|
|
|
my $s = shift; |
|
1433
|
|
|
|
|
|
|
substr $s->{buf}, 0, $s->{len}; |
|
1434
|
|
|
|
|
|
|
} |
|
1435
|
|
|
|
|
|
|
|
|
1436
|
|
|
|
|
|
|
sub string_pack2 { |
|
1437
|
|
|
|
|
|
|
my $s = shift; |
|
1438
|
|
|
|
|
|
|
return { |
|
1439
|
|
|
|
|
|
|
len => length $s, |
|
1440
|
|
|
|
|
|
|
buf => $s, |
|
1441
|
|
|
|
|
|
|
} |
|
1442
|
|
|
|
|
|
|
} |
|
1443
|
|
|
|
|
|
|
|
|
1444
|
|
|
|
|
|
|
$c->tag('String', Hooks => { pack => \&string_pack2, |
|
1445
|
|
|
|
|
|
|
unpack => \&string_unpack2 }); |
|
1446
|
|
|
|
|
|
|
|
|
1447
|
|
|
|
|
|
|
This will be exactly equivalent to the old code, but faster and |
|
1448
|
|
|
|
|
|
|
probably even much easier to understand. |
|
1449
|
|
|
|
|
|
|
|
|
1450
|
|
|
|
|
|
|
But hooks are even more powerful. You can customize the arguments |
|
1451
|
|
|
|
|
|
|
that are passed to your hooks and you can use L<C<arg>|/"arg"> to |
|
1452
|
|
|
|
|
|
|
pass certain special arguments, such as the name of the type that |
|
1453
|
|
|
|
|
|
|
is currently being processed by the hook. |
|
1454
|
|
|
|
|
|
|
|
|
1455
|
|
|
|
|
|
|
The following example shows how it is easily possible to peek into |
|
1456
|
|
|
|
|
|
|
the perl internals using hooks. |
|
1457
|
|
|
|
|
|
|
|
|
1458
|
|
|
|
|
|
|
use Config; |
|
1459
|
|
|
|
|
|
|
|
|
1460
|
|
|
|
|
|
|
$c = new Convert::Binary::C %CC, OrderMembers => 1; |
|
1461
|
|
|
|
|
|
|
$c->Include(["$Config{archlib}/CORE", @{$c->Include}]); |
|
1462
|
|
|
|
|
|
|
$c->parse(<<ENDC); |
|
1463
|
|
|
|
|
|
|
#include "EXTERN.h" |
|
1464
|
|
|
|
|
|
|
#include "perl.h" |
|
1465
|
|
|
|
|
|
|
ENDC |
|
1466
|
|
|
|
|
|
|
|
|
1467
|
|
|
|
|
|
|
$c->tag($_, Hooks => { unpack_ptr => [\&unpack_ptr, |
|
1468
|
|
|
|
|
|
|
$c->arg(qw(SELF TYPE DATA))] }) |
|
1469
|
|
|
|
|
|
|
for qw( XPVAV XPVHV MAGIC MGVTBL HV ); |
|
1470
|
|
|
|
|
|
|
|
|
1471
|
|
|
|
|
|
|
First, we add the perl core include path and parse F<perl.h>. Then, |
|
1472
|
|
|
|
|
|
|
we add an C<unpack_ptr> hook for a couple of the internal data types. |
|
1473
|
|
|
|
|
|
|
|
|
1474
|
|
|
|
|
|
|
The C<unpack_ptr> and C<pack_ptr> hooks are called whenever a pointer |
|
1475
|
|
|
|
|
|
|
to a certain data structure is processed. This is by far the most |
|
1476
|
|
|
|
|
|
|
experimental part of the hooks feature, as this includes B<any> kind |
|
1477
|
|
|
|
|
|
|
of pointer. There's no way for the hook to know the difference between |
|
1478
|
|
|
|
|
|
|
a plain pointer, or a pointer to a pointer, or a pointer to an array |
|
1479
|
|
|
|
|
|
|
(this is because the difference doesn't matter anywhere else in |
|
1480
|
|
|
|
|
|
|
Convert::Binary::C). |
|
1481
|
|
|
|
|
|
|
|
|
1482
|
|
|
|
|
|
|
But the hook above makes use of another very interesting feature: It |
|
1483
|
|
|
|
|
|
|
uses L<C<arg>|/"arg"> to pass special arguments to the hook subroutine. |
|
1484
|
|
|
|
|
|
|
Usually, the hook subroutine is simply passed a single data argument. |
|
1485
|
|
|
|
|
|
|
But using the above definition, it'll get a reference to the calling |
|
1486
|
|
|
|
|
|
|
object (C<SELF>), the name of the type being processed (C<TYPE>) and |
|
1487
|
|
|
|
|
|
|
the data (C<DATA>). |
|
1488
|
|
|
|
|
|
|
|
|
1489
|
|
|
|
|
|
|
But how does our hook look like? |
|
1490
|
|
|
|
|
|
|
|
|
1491
|
|
|
|
|
|
|
sub unpack_ptr { |
|
1492
|
|
|
|
|
|
|
my($self, $type, $ptr) = @_; |
|
1493
|
|
|
|
|
|
|
$ptr or return '<NULL>'; |
|
1494
|
|
|
|
|
|
|
my $size = $self->sizeof($type); |
|
1495
|
|
|
|
|
|
|
$self->unpack($type, unpack("P$size", pack('I', $ptr))); |
|
1496
|
|
|
|
|
|
|
} |
|
1497
|
|
|
|
|
|
|
|
|
1498
|
|
|
|
|
|
|
As you can see, the hook is rather simple. First, it receives the |
|
1499
|
|
|
|
|
|
|
arguments mentioned above. It performs a quick check if the pointer |
|
1500
|
|
|
|
|
|
|
is C<NULL> and shouldn't be processed any further. Next, it determines |
|
1501
|
|
|
|
|
|
|
the size of the type being processed. And finally, it'll just use |
|
1502
|
|
|
|
|
|
|
the C<P>I<n> unpack template to read from that memory location and |
|
1503
|
|
|
|
|
|
|
recursively call L<C<unpack>|/"unpack"> to unpack the type. (And yes, |
|
1504
|
|
|
|
|
|
|
this may of course again call other hooks.) |
|
1505
|
|
|
|
|
|
|
|
|
1506
|
|
|
|
|
|
|
Now, let's test that: |
|
1507
|
|
|
|
|
|
|
|
|
1508
|
|
|
|
|
|
|
my $ref = bless ["Boo!"], "Foo::Bar"; |
|
1509
|
|
|
|
|
|
|
my $ptr = hex(("$ref" =~ /\(0x([[:xdigit:]]+)\)$/)[0]); |
|
1510
|
|
|
|
|
|
|
|
|
1511
|
|
|
|
|
|
|
print Dumper(unpack_ptr($c, 'AV', $ptr)); |
|
1512
|
|
|
|
|
|
|
|
|
1513
|
|
|
|
|
|
|
Just for the fun of it, we create a blessed array reference. But how |
|
1514
|
|
|
|
|
|
|
do we get a pointer to the corresponding C<AV>? This is rather easy, |
|
1515
|
|
|
|
|
|
|
as the address of the C<AV> is just the hex value that appears when |
|
1516
|
|
|
|
|
|
|
using the array reference in string context. So we just grab that and |
|
1517
|
|
|
|
|
|
|
turn it into decimal. All that's left to do is just call our hook, |
|
1518
|
|
|
|
|
|
|
as it can already handle C<AV> pointers. And this is what we get: |
|
1519
|
|
|
|
|
|
|
|
|
1520
|
|
|
|
|
|
|
$VAR1 = { |
|
1521
|
|
|
|
|
|
|
'sv_any' => { |
|
1522
|
|
|
|
|
|
|
'xnv_u' => { |
|
1523
|
|
|
|
|
|
|
'xnv_nv' => '0', |
|
1524
|
|
|
|
|
|
|
'xgv_stash' => '<NULL>' |
|
1525
|
|
|
|
|
|
|
}, |
|
1526
|
|
|
|
|
|
|
'xav_fill' => 0, |
|
1527
|
|
|
|
|
|
|
'xav_max' => 0, |
|
1528
|
|
|
|
|
|
|
'xiv_u' => { |
|
1529
|
|
|
|
|
|
|
'xivu_iv' => 140312788, |
|
1530
|
|
|
|
|
|
|
'xivu_uv' => 140312788, |
|
1531
|
|
|
|
|
|
|
'xivu_p1' => 140312788, |
|
1532
|
|
|
|
|
|
|
'xivu_i32' => 140312788, |
|
1533
|
|
|
|
|
|
|
'xivu_namehek' => 140312788 |
|
1534
|
|
|
|
|
|
|
}, |
|
1535
|
|
|
|
|
|
|
'xmg_u' => { |
|
1536
|
|
|
|
|
|
|
'xmg_magic' => '<NULL>', |
|
1537
|
|
|
|
|
|
|
'xmg_ourstash' => '<NULL>' |
|
1538
|
|
|
|
|
|
|
}, |
|
1539
|
|
|
|
|
|
|
'xmg_stash' => { |
|
1540
|
|
|
|
|
|
|
'sv_any' => { |
|
1541
|
|
|
|
|
|
|
'xnv_u' => { |
|
1542
|
|
|
|
|
|
|
'xnv_nv' => '0', |
|
1543
|
|
|
|
|
|
|
'xgv_stash' => '<NULL>' |
|
1544
|
|
|
|
|
|
|
}, |
|
1545
|
|
|
|
|
|
|
'xhv_fill' => 2, |
|
1546
|
|
|
|
|
|
|
'xhv_max' => 7, |
|
1547
|
|
|
|
|
|
|
'xiv_u' => { |
|
1548
|
|
|
|
|
|
|
'xivu_iv' => 2, |
|
1549
|
|
|
|
|
|
|
'xivu_uv' => 2, |
|
1550
|
|
|
|
|
|
|
'xivu_p1' => 2, |
|
1551
|
|
|
|
|
|
|
'xivu_i32' => 2, |
|
1552
|
|
|
|
|
|
|
'xivu_namehek' => 2 |
|
1553
|
|
|
|
|
|
|
}, |
|
1554
|
|
|
|
|
|
|
'xmg_u' => { |
|
1555
|
|
|
|
|
|
|
'xmg_magic' => { |
|
1556
|
|
|
|
|
|
|
'mg_moremagic' => '<NULL>', |
|
1557
|
|
|
|
|
|
|
'mg_virtual' => { |
|
1558
|
|
|
|
|
|
|
'svt_get' => 0, |
|
1559
|
|
|
|
|
|
|
'svt_set' => 0, |
|
1560
|
|
|
|
|
|
|
'svt_len' => 0, |
|
1561
|
|
|
|
|
|
|
'svt_clear' => 0, |
|
1562
|
|
|
|
|
|
|
'svt_free' => 136674986, |
|
1563
|
|
|
|
|
|
|
'svt_copy' => 0, |
|
1564
|
|
|
|
|
|
|
'svt_dup' => 0, |
|
1565
|
|
|
|
|
|
|
'svt_local' => 0 |
|
1566
|
|
|
|
|
|
|
}, |
|
1567
|
|
|
|
|
|
|
'mg_private' => 0, |
|
1568
|
|
|
|
|
|
|
'mg_type' => 99, |
|
1569
|
|
|
|
|
|
|
'mg_flags' => 0, |
|
1570
|
|
|
|
|
|
|
'mg_obj' => 0, |
|
1571
|
|
|
|
|
|
|
'mg_ptr' => 139425604, |
|
1572
|
|
|
|
|
|
|
'mg_len' => 12 |
|
1573
|
|
|
|
|
|
|
}, |
|
1574
|
|
|
|
|
|
|
'xmg_ourstash' => { |
|
1575
|
|
|
|
|
|
|
'sv_any' => '<NULL>', |
|
1576
|
|
|
|
|
|
|
'sv_refcnt' => 137217696, |
|
1577
|
|
|
|
|
|
|
'sv_flags' => 6488064, |
|
1578
|
|
|
|
|
|
|
'sv_u' => { |
|
1579
|
|
|
|
|
|
|
'svu_iv' => '598828409405046784', |
|
1580
|
|
|
|
|
|
|
'svu_uv' => '598828409405046784', |
|
1581
|
|
|
|
|
|
|
'svu_rv' => 0, |
|
1582
|
|
|
|
|
|
|
'svu_pv' => 0, |
|
1583
|
|
|
|
|
|
|
'svu_array' => 0, |
|
1584
|
|
|
|
|
|
|
'svu_hash' => 0, |
|
1585
|
|
|
|
|
|
|
'svu_gp' => 0 |
|
1586
|
|
|
|
|
|
|
} |
|
1587
|
|
|
|
|
|
|
} |
|
1588
|
|
|
|
|
|
|
}, |
|
1589
|
|
|
|
|
|
|
'xmg_stash' => '<NULL>' |
|
1590
|
|
|
|
|
|
|
}, |
|
1591
|
|
|
|
|
|
|
'sv_refcnt' => 2, |
|
1592
|
|
|
|
|
|
|
'sv_flags' => 578813964, |
|
1593
|
|
|
|
|
|
|
'sv_u' => { |
|
1594
|
|
|
|
|
|
|
'svu_iv' => '2314885530418902580', |
|
1595
|
|
|
|
|
|
|
'svu_uv' => '2314885530418902580', |
|
1596
|
|
|
|
|
|
|
'svu_rv' => 139425332, |
|
1597
|
|
|
|
|
|
|
'svu_pv' => 139425332, |
|
1598
|
|
|
|
|
|
|
'svu_array' => 139425332, |
|
1599
|
|
|
|
|
|
|
'svu_hash' => 139425332, |
|
1600
|
|
|
|
|
|
|
'svu_gp' => 139425332 |
|
1601
|
|
|
|
|
|
|
} |
|
1602
|
|
|
|
|
|
|
} |
|
1603
|
|
|
|
|
|
|
}, |
|
1604
|
|
|
|
|
|
|
'sv_refcnt' => 1, |
|
1605
|
|
|
|
|
|
|
'sv_flags' => 1074790411, |
|
1606
|
|
|
|
|
|
|
'sv_u' => { |
|
1607
|
|
|
|
|
|
|
'svu_iv' => '3689628089976684756', |
|
1608
|
|
|
|
|
|
|
'svu_uv' => '3689628089976684756', |
|
1609
|
|
|
|
|
|
|
'svu_rv' => 140312788, |
|
1610
|
|
|
|
|
|
|
'svu_pv' => 140312788, |
|
1611
|
|
|
|
|
|
|
'svu_array' => 140312788, |
|
1612
|
|
|
|
|
|
|
'svu_hash' => 140312788, |
|
1613
|
|
|
|
|
|
|
'svu_gp' => 140312788 |
|
1614
|
|
|
|
|
|
|
} |
|
1615
|
|
|
|
|
|
|
}; |
|
1616
|
|
|
|
|
|
|
|
|
1617
|
|
|
|
|
|
|
Even though it is rather easy to do such stuff using C<unpack_ptr> hooks, |
|
1618
|
|
|
|
|
|
|
you should really know what you're doing and do it with extreme care |
|
1619
|
|
|
|
|
|
|
because of the limitations mentioned above. It's really easy to run into |
|
1620
|
|
|
|
|
|
|
segmentation faults when you're dereferencing pointers that point to |
|
1621
|
|
|
|
|
|
|
memory which you don't own. |
|
1622
|
|
|
|
|
|
|
|
|
1623
|
|
|
|
|
|
|
=head3 Performance |
|
1624
|
|
|
|
|
|
|
|
|
1625
|
|
|
|
|
|
|
Using hooks isn't for free. In performance-critical applications |
|
1626
|
|
|
|
|
|
|
you have to keep in mind that hooks are actually perl subroutines |
|
1627
|
|
|
|
|
|
|
and that they are called once for every value of a registered |
|
1628
|
|
|
|
|
|
|
type that is being packed or unpacked. If only about 10% of the |
|
1629
|
|
|
|
|
|
|
values require hooks to be called, you'll hardly notice the |
|
1630
|
|
|
|
|
|
|
difference (if your hooks are implemented efficiently, that is). |
|
1631
|
|
|
|
|
|
|
But if all values would require hooks to be called, that alone |
|
1632
|
|
|
|
|
|
|
could easily make packing and unpacking very slow. |
|
1633
|
|
|
|
|
|
|
|
|
1634
|
|
|
|
|
|
|
=head2 Tag Order |
|
1635
|
|
|
|
|
|
|
|
|
1636
|
|
|
|
|
|
|
Since it is possible to attach multiple tags to a single type, |
|
1637
|
|
|
|
|
|
|
the order in which the tags are processed is important. Here's |
|
1638
|
|
|
|
|
|
|
a small table that shows the processing order. |
|
1639
|
|
|
|
|
|
|
|
|
1640
|
|
|
|
|
|
|
pack unpack |
|
1641
|
|
|
|
|
|
|
--------------------- |
|
1642
|
|
|
|
|
|
|
Hooks Format |
|
1643
|
|
|
|
|
|
|
Format ByteOrder |
|
1644
|
|
|
|
|
|
|
ByteOrder Hooks |
|
1645
|
|
|
|
|
|
|
|
|
1646
|
|
|
|
|
|
|
As a general rule, the L<C<Hooks>|/"The Hooks Tag"> tag is always |
|
1647
|
|
|
|
|
|
|
the first thing processed when packing data, and the last thing |
|
1648
|
|
|
|
|
|
|
processed when unpacking data. |
|
1649
|
|
|
|
|
|
|
|
|
1650
|
|
|
|
|
|
|
The L<C<Format>|/"The Format Tag"> and L<C<ByteOrder>|/"The ByteOrder Tag"> tags |
|
1651
|
|
|
|
|
|
|
are exclusive, but when both are given the L<C<Format>|/"The Format Tag"> tag |
|
1652
|
|
|
|
|
|
|
wins. |
|
1653
|
|
|
|
|
|
|
|
|
1654
|
|
|
|
|
|
|
=head1 METHODS |
|
1655
|
|
|
|
|
|
|
|
|
1656
|
|
|
|
|
|
|
=head2 new |
|
1657
|
|
|
|
|
|
|
|
|
1658
|
|
|
|
|
|
|
=over 8 |
|
1659
|
|
|
|
|
|
|
|
|
1660
|
|
|
|
|
|
|
=item C<new> |
|
1661
|
|
|
|
|
|
|
|
|
1662
|
|
|
|
|
|
|
=item C<new> OPTION1 =E<gt> VALUE1, OPTION2 =E<gt> VALUE2, ... |
|
1663
|
|
|
|
|
|
|
|
|
1664
|
|
|
|
|
|
|
The constructor is used to create a new Convert::Binary::C object. |
|
1665
|
|
|
|
|
|
|
You can simply use |
|
1666
|
|
|
|
|
|
|
|
|
1667
|
|
|
|
|
|
|
$c = new Convert::Binary::C; |
|
1668
|
|
|
|
|
|
|
|
|
1669
|
|
|
|
|
|
|
without additional arguments to create an object, or you can |
|
1670
|
|
|
|
|
|
|
optionally pass any arguments to the constructor that are |
|
1671
|
|
|
|
|
|
|
described for the L<C<configure>|/"configure"> method. |
|
1672
|
|
|
|
|
|
|
|
|
1673
|
|
|
|
|
|
|
=back |
|
1674
|
|
|
|
|
|
|
|
|
1675
|
|
|
|
|
|
|
=head2 configure |
|
1676
|
|
|
|
|
|
|
|
|
1677
|
|
|
|
|
|
|
=over 8 |
|
1678
|
|
|
|
|
|
|
|
|
1679
|
|
|
|
|
|
|
=item C<configure> |
|
1680
|
|
|
|
|
|
|
|
|
1681
|
|
|
|
|
|
|
=item C<configure> OPTION |
|
1682
|
|
|
|
|
|
|
|
|
1683
|
|
|
|
|
|
|
=item C<configure> OPTION1 =E<gt> VALUE1, OPTION2 =E<gt> VALUE2, ... |
|
1684
|
|
|
|
|
|
|
|
|
1685
|
|
|
|
|
|
|
This method can be used to configure an existing Convert::Binary::C |
|
1686
|
|
|
|
|
|
|
object or to retrieve its current configuration. |
|
1687
|
|
|
|
|
|
|
|
|
1688
|
|
|
|
|
|
|
To configure the object, the list of options consists of key |
|
1689
|
|
|
|
|
|
|
and value pairs and must therefore contain an even number of |
|
1690
|
|
|
|
|
|
|
elements. L<C<configure>|/"configure"> (and also L<C<new>|/"new"> if |
|
1691
|
|
|
|
|
|
|
used with configuration options) will throw an exception if you |
|
1692
|
|
|
|
|
|
|
pass an odd number of elements. Configuration will normally look |
|
1693
|
|
|
|
|
|
|
like this: |
|
1694
|
|
|
|
|
|
|
|
|
1695
|
|
|
|
|
|
|
$c->configure(ByteOrder => 'BigEndian', IntSize => 2); |
|
1696
|
|
|
|
|
|
|
|
|
1697
|
|
|
|
|
|
|
To retrieve the current value of a configuration option, you |
|
1698
|
|
|
|
|
|
|
must pass a single argument to L<C<configure>|/"configure"> that |
|
1699
|
|
|
|
|
|
|
holds the name of the option, just like |
|
1700
|
|
|
|
|
|
|
|
|
1701
|
|
|
|
|
|
|
$order = $c->configure('ByteOrder'); |
|
1702
|
|
|
|
|
|
|
|
|
1703
|
|
|
|
|
|
|
If you want to get the values of all configuration options at |
|
1704
|
|
|
|
|
|
|
once, you can call L<C<configure>|/"configure"> without any |
|
1705
|
|
|
|
|
|
|
arguments and it will return a reference to a hash table that |
|
1706
|
|
|
|
|
|
|
holds the whole object configuration. This can be conveniently |
|
1707
|
|
|
|
|
|
|
used with the L<Data::Dumper|Data::Dumper> module, for example: |
|
1708
|
|
|
|
|
|
|
|
|
1709
|
|
|
|
|
|
|
use Convert::Binary::C; |
|
1710
|
|
|
|
|
|
|
use Data::Dumper; |
|
1711
|
|
|
|
|
|
|
|
|
1712
|
|
|
|
|
|
|
$c = new Convert::Binary::C Define => ['DEBUGGING', 'FOO=123'], |
|
1713
|
|
|
|
|
|
|
Include => ['/usr/include']; |
|
1714
|
|
|
|
|
|
|
|
|
1715
|
|
|
|
|
|
|
print Dumper($c->configure); |
|
1716
|
|
|
|
|
|
|
|
|
1717
|
|
|
|
|
|
|
Which will print something like this: |
|
1718
|
|
|
|
|
|
|
|
|
1719
|
|
|
|
|
|
|
$VAR1 = { |
|
1720
|
|
|
|
|
|
|
'Define' => [ |
|
1721
|
|
|
|
|
|
|
'DEBUGGING', |
|
1722
|
|
|
|
|
|
|
'FOO=123' |
|
1723
|
|
|
|
|
|
|
], |
|
1724
|
|
|
|
|
|
|
'ByteOrder' => 'LittleEndian', |
|
1725
|
|
|
|
|
|
|
'LongSize' => 4, |
|
1726
|
|
|
|
|
|
|
'IntSize' => 4, |
|
1727
|
|
|
|
|
|
|
'ShortSize' => 2, |
|
1728
|
|
|
|
|
|
|
'HasMacroVAARGS' => 1, |
|
1729
|
|
|
|
|
|
|
'Assert' => [], |
|
1730
|
|
|
|
|
|
|
'UnsignedChars' => 0, |
|
1731
|
|
|
|
|
|
|
'DoubleSize' => 8, |
|
1732
|
|
|
|
|
|
|
'CharSize' => 1, |
|
1733
|
|
|
|
|
|
|
'EnumType' => 'Integer', |
|
1734
|
|
|
|
|
|
|
'PointerSize' => 4, |
|
1735
|
|
|
|
|
|
|
'EnumSize' => 4, |
|
1736
|
|
|
|
|
|
|
'DisabledKeywords' => [], |
|
1737
|
|
|
|
|
|
|
'FloatSize' => 4, |
|
1738
|
|
|
|
|
|
|
'Alignment' => 1, |
|
1739
|
|
|
|
|
|
|
'LongLongSize' => 8, |
|
1740
|
|
|
|
|
|
|
'LongDoubleSize' => 12, |
|
1741
|
|
|
|
|
|
|
'KeywordMap' => {}, |
|
1742
|
|
|
|
|
|
|
'Include' => [ |
|
1743
|
|
|
|
|
|
|
'/usr/include' |
|
1744
|
|
|
|
|
|
|
], |
|
1745
|
|
|
|
|
|
|
'HasCPPComments' => 1, |
|
1746
|
|
|
|
|
|
|
'Bitfields' => { |
|
1747
|
|
|
|
|
|
|
'Engine' => 'Generic' |
|
1748
|
|
|
|
|
|
|
}, |
|
1749
|
|
|
|
|
|
|
'UnsignedBitfields' => 0, |
|
1750
|
|
|
|
|
|
|
'Warnings' => 0, |
|
1751
|
|
|
|
|
|
|
'CompoundAlignment' => 1, |
|
1752
|
|
|
|
|
|
|
'OrderMembers' => 0 |
|
1753
|
|
|
|
|
|
|
}; |
|
1754
|
|
|
|
|
|
|
|
|
1755
|
|
|
|
|
|
|
Since you may not always want to write a L<C<configure>|/"configure"> call |
|
1756
|
|
|
|
|
|
|
when you only want to change a single configuration item, you can |
|
1757
|
|
|
|
|
|
|
use any configuration option name as a method name, like: |
|
1758
|
|
|
|
|
|
|
|
|
1759
|
|
|
|
|
|
|
$c->ByteOrder('LittleEndian') if $c->IntSize < 4; |
|
1760
|
|
|
|
|
|
|
|
|
1761
|
|
|
|
|
|
|
(Yes, the example doesn't make very much sense... ;-) |
|
1762
|
|
|
|
|
|
|
|
|
1763
|
|
|
|
|
|
|
However, you should keep in mind that configuration methods |
|
1764
|
|
|
|
|
|
|
that can take lists (namely C<Include>, C<Define> and C<Assert>, |
|
1765
|
|
|
|
|
|
|
but not C<DisabledKeywords>) may behave slightly different than |
|
1766
|
|
|
|
|
|
|
their L<C<configure>|/"configure"> equivalent. |
|
1767
|
|
|
|
|
|
|
If you pass these methods a single argument that is an array |
|
1768
|
|
|
|
|
|
|
reference, the current list will be B<replaced> by the new one, |
|
1769
|
|
|
|
|
|
|
which is just the behaviour of the |
|
1770
|
|
|
|
|
|
|
corresponding L<C<configure>|/"configure"> call. |
|
1771
|
|
|
|
|
|
|
So the following are equivalent: |
|
1772
|
|
|
|
|
|
|
|
|
1773
|
|
|
|
|
|
|
$c->configure(Define => ['foo', 'bar=123']); |
|
1774
|
|
|
|
|
|
|
$c->Define(['foo', 'bar=123']); |
|
1775
|
|
|
|
|
|
|
|
|
1776
|
|
|
|
|
|
|
But if you pass a list of strings instead of an array reference |
|
1777
|
|
|
|
|
|
|
(which cannot be done when using L<C<configure>|/"configure">), |
|
1778
|
|
|
|
|
|
|
the new list items are B<appended> to the current list, so |
|
1779
|
|
|
|
|
|
|
|
|
1780
|
|
|
|
|
|
|
$c = new Convert::Binary::C Include => ['/include']; |
|
1781
|
|
|
|
|
|
|
$c->Include('/usr/include', '/usr/local/include'); |
|
1782
|
|
|
|
|
|
|
print Dumper($c->Include); |
|
1783
|
|
|
|
|
|
|
|
|
1784
|
|
|
|
|
|
|
$c->Include(['/usr/local/include']); |
|
1785
|
|
|
|
|
|
|
print Dumper($c->Include); |
|
1786
|
|
|
|
|
|
|
|
|
1787
|
|
|
|
|
|
|
will first print all three include paths, but finally |
|
1788
|
|
|
|
|
|
|
only C</usr/local/include> will be configured: |
|
1789
|
|
|
|
|
|
|
|
|
1790
|
|
|
|
|
|
|
$VAR1 = [ |
|
1791
|
|
|
|
|
|
|
'/include', |
|
1792
|
|
|
|
|
|
|
'/usr/include', |
|
1793
|
|
|
|
|
|
|
'/usr/local/include' |
|
1794
|
|
|
|
|
|
|
]; |
|
1795
|
|
|
|
|
|
|
$VAR1 = [ |
|
1796
|
|
|
|
|
|
|
'/usr/local/include' |
|
1797
|
|
|
|
|
|
|
]; |
|
1798
|
|
|
|
|
|
|
|
|
1799
|
|
|
|
|
|
|
Furthermore, configuration methods can be chained together, |
|
1800
|
|
|
|
|
|
|
as they return a reference to their object if called as a |
|
1801
|
|
|
|
|
|
|
set method. So, if you like, you can configure your object |
|
1802
|
|
|
|
|
|
|
like this: |
|
1803
|
|
|
|
|
|
|
|
|
1804
|
|
|
|
|
|
|
$c = Convert::Binary::C->new(IntSize => 4) |
|
1805
|
|
|
|
|
|
|
->Define(qw( __DEBUG__ DB_LEVEL=3 )) |
|
1806
|
|
|
|
|
|
|
->ByteOrder('BigEndian'); |
|
1807
|
|
|
|
|
|
|
|
|
1808
|
|
|
|
|
|
|
$c->configure(EnumType => 'Both', Alignment => 4) |
|
1809
|
|
|
|
|
|
|
->Include('/usr/include', '/usr/local/include'); |
|
1810
|
|
|
|
|
|
|
|
|
1811
|
|
|
|
|
|
|
In the example above, C<qw( ... )> is the word list quoting |
|
1812
|
|
|
|
|
|
|
operator. It returns a list of all non-whitespace sequences, |
|
1813
|
|
|
|
|
|
|
and is especially useful for configuring preprocessor defines |
|
1814
|
|
|
|
|
|
|
or assertions. The following assignments are equivalent: |
|
1815
|
|
|
|
|
|
|
|
|
1816
|
|
|
|
|
|
|
@array = ('one', 'two', 'three'); |
|
1817
|
|
|
|
|
|
|
@array = qw(one two three); |
|
1818
|
|
|
|
|
|
|
|
|
1819
|
|
|
|
|
|
|
You can configure the following options. Unknown options, as well |
|
1820
|
|
|
|
|
|
|
as invalid values for an option, will cause the object to throw |
|
1821
|
|
|
|
|
|
|
exceptions. |
|
1822
|
|
|
|
|
|
|
|
|
1823
|
|
|
|
|
|
|
=over 4 |
|
1824
|
|
|
|
|
|
|
|
|
1825
|
|
|
|
|
|
|
=item C<IntSize> =E<gt> 0 | 1 | 2 | 4 | 8 |
|
1826
|
|
|
|
|
|
|
|
|
1827
|
|
|
|
|
|
|
Set the number of bytes that are occupied by an integer. This is |
|
1828
|
|
|
|
|
|
|
in most cases 2 or 4. If you set it to zero, the size of an |
|
1829
|
|
|
|
|
|
|
integer on the host system will be used. This is also the |
|
1830
|
|
|
|
|
|
|
default unless overridden by C<CBC_DEFAULT_INT_SIZE> at compile time. |
|
1831
|
|
|
|
|
|
|
|
|
1832
|
|
|
|
|
|
|
=item C<CharSize> =E<gt> 0 | 1 | 2 | 4 | 8 |
|
1833
|
|
|
|
|
|
|
|
|
1834
|
|
|
|
|
|
|
Set the number of bytes that are occupied by a C<char>. |
|
1835
|
|
|
|
|
|
|
This rarely needs to be changed, except for some platforms |
|
1836
|
|
|
|
|
|
|
that don't care about bytes, for example DSPs. |
|
1837
|
|
|
|
|
|
|
If you set this to zero, the size of a C<char> on the host |
|
1838
|
|
|
|
|
|
|
system will be used. This is also the default unless |
|
1839
|
|
|
|
|
|
|
overridden by C<CBC_DEFAULT_CHAR_SIZE> at compile time. |
|
1840
|
|
|
|
|
|
|
|
|
1841
|
|
|
|
|
|
|
=item C<ShortSize> =E<gt> 0 | 1 | 2 | 4 | 8 |
|
1842
|
|
|
|
|
|
|
|
|
1843
|
|
|
|
|
|
|
Set the number of bytes that are occupied by a short integer. |
|
1844
|
|
|
|
|
|
|
Although integers explicitly declared as C<short> should be |
|
1845
|
|
|
|
|
|
|
always 16 bit, there are compilers that make a short |
|
1846
|
|
|
|
|
|
|
8 bit wide. If you set it to zero, the size of a short |
|
1847
|
|
|
|
|
|
|
integer on the host system will be used. This is also the |
|
1848
|
|
|
|
|
|
|
default unless overridden by C<CBC_DEFAULT_SHORT_SIZE> at compile |
|
1849
|
|
|
|
|
|
|
time. |
|
1850
|
|
|
|
|
|
|
|
|
1851
|
|
|
|
|
|
|
=item C<LongSize> =E<gt> 0 | 1 | 2 | 4 | 8 |
|
1852
|
|
|
|
|
|
|
|
|
1853
|
|
|
|
|
|
|
Set the number of bytes that are occupied by a long integer. |
|
1854
|
|
|
|
|
|
|
If set to zero, the size of a long integer on the host system |
|
1855
|
|
|
|
|
|
|
will be used. This is also the default unless overridden |
|
1856
|
|
|
|
|
|
|
by C<CBC_DEFAULT_LONG_SIZE> at compile time. |
|
1857
|
|
|
|
|
|
|
|
|
1858
|
|
|
|
|
|
|
=item C<LongLongSize> =E<gt> 0 | 1 | 2 | 4 | 8 |
|
1859
|
|
|
|
|
|
|
|
|
1860
|
|
|
|
|
|
|
Set the number of bytes that are occupied by a long long |
|
1861
|
|
|
|
|
|
|
integer. If set to zero, the size of a long long integer |
|
1862
|
|
|
|
|
|
|
on the host system, or 8, will be used. This is also the |
|
1863
|
|
|
|
|
|
|
default unless overridden by C<CBC_DEFAULT_LONG_LONG_SIZE> at |
|
1864
|
|
|
|
|
|
|
compile time. |
|
1865
|
|
|
|
|
|
|
|
|
1866
|
|
|
|
|
|
|
=item C<FloatSize> =E<gt> 0 | 1 | 2 | 4 | 8 | 12 | 16 |
|
1867
|
|
|
|
|
|
|
|
|
1868
|
|
|
|
|
|
|
Set the number of bytes that are occupied by a single |
|
1869
|
|
|
|
|
|
|
precision floating point value. |
|
1870
|
|
|
|
|
|
|
If you set it to zero, the size of a C<float> on the |
|
1871
|
|
|
|
|
|
|
host system will be used. This is also the default unless |
|
1872
|
|
|
|
|
|
|
overridden by C<CBC_DEFAULT_FLOAT_SIZE> at compile time. |
|
1873
|
|
|
|
|
|
|
For details on floating point support, |
|
1874
|
|
|
|
|
|
|
see L<"FLOATING POINT VALUES">. |
|
1875
|
|
|
|
|
|
|
|
|
1876
|
|
|
|
|
|
|
=item C<DoubleSize> =E<gt> 0 | 1 | 2 | 4 | 8 | 12 | 16 |
|
1877
|
|
|
|
|
|
|
|
|
1878
|
|
|
|
|
|
|
Set the number of bytes that are occupied by a double |
|
1879
|
|
|
|
|
|
|
precision floating point value. |
|
1880
|
|
|
|
|
|
|
If you set it to zero, the size of a C<double> on the |
|
1881
|
|
|
|
|
|
|
host system will be used. This is also the default unless |
|
1882
|
|
|
|
|
|
|
overridden by C<CBC_DEFAULT_DOUBLE_SIZE> at compile time. |
|
1883
|
|
|
|
|
|
|
For details on floating point support, |
|
1884
|
|
|
|
|
|
|
see L<"FLOATING POINT VALUES">. |
|
1885
|
|
|
|
|
|
|
|
|
1886
|
|
|
|
|
|
|
=item C<LongDoubleSize> =E<gt> 0 | 1 | 2 | 4 | 8 | 12 | 16 |
|
1887
|
|
|
|
|
|
|
|
|
1888
|
|
|
|
|
|
|
Set the number of bytes that are occupied by a double |
|
1889
|
|
|
|
|
|
|
precision floating point value. |
|
1890
|
|
|
|
|
|
|
If you set it to zero, the size of a C<long double> on |
|
1891
|
|
|
|
|
|
|
the host system, or 12 will be used. This is also the |
|
1892
|
|
|
|
|
|
|
default unless overridden by C<CBC_DEFAULT_LONG_DOUBLE_SIZE> at compile |
|
1893
|
|
|
|
|
|
|
time. For details on floating point support, |
|
1894
|
|
|
|
|
|
|
see L<"FLOATING POINT VALUES">. |
|
1895
|
|
|
|
|
|
|
|
|
1896
|
|
|
|
|
|
|
=item C<PointerSize> =E<gt> 0 | 1 | 2 | 4 | 8 |
|
1897
|
|
|
|
|
|
|
|
|
1898
|
|
|
|
|
|
|
Set the number of bytes that are occupied by a pointer. This is |
|
1899
|
|
|
|
|
|
|
in most cases 2 or 4. If you set it to zero, the size of a |
|
1900
|
|
|
|
|
|
|
pointer on the host system will be used. This is also the |
|
1901
|
|
|
|
|
|
|
default unless overridden by C<CBC_DEFAULT_PTR_SIZE> at compile time. |
|
1902
|
|
|
|
|
|
|
|
|
1903
|
|
|
|
|
|
|
=item C<EnumSize> =E<gt> -1 | 0 | 1 | 2 | 4 | 8 |
|
1904
|
|
|
|
|
|
|
|
|
1905
|
|
|
|
|
|
|
Set the number of bytes that are occupied by an enumeration type. |
|
1906
|
|
|
|
|
|
|
On most systems, this is equal to the size of an integer, |
|
1907
|
|
|
|
|
|
|
which is also the default. However, for some compilers, the |
|
1908
|
|
|
|
|
|
|
size of an enumeration type depends on the size occupied by the |
|
1909
|
|
|
|
|
|
|
largest enumerator. So the size may vary between 1 and 8. If you |
|
1910
|
|
|
|
|
|
|
have |
|
1911
|
|
|
|
|
|
|
|
|
1912
|
|
|
|
|
|
|
enum foo { |
|
1913
|
|
|
|
|
|
|
ONE = 100, TWO = 200 |
|
1914
|
|