Read/write buffer class
use Data::Buffer;
my $buffer = Data::Buffer->new;
## Add a 32-bit integer. $buffer->put_int32(10932930);
## Get it back. my $int = $buffer->get_int32;
Data::Buffer implements a low-level binary buffer in which you can get and put integers, strings, and other data. Internally the implementation is based on \*(C`pack\*(C' and \*(C`unpack\*(C', such that Data::Buffer is really a layer on top of those built-in functions.
All of the get_* and put_* methods respect the internal offset state in the buffer object. This means that you should read data out of the buffer in the same order that you put it in. For example:
$buf->put_int16(24); $buf->put_int32(1233455); $buf->put_int16(99);
$buf->get_int16; # 24 $buf->get_int32; # 1233455 $buf->get_int16; # 99
Of course, this assumes that you know the order of the data items in the buffer. If your setup is such that your sending and receiving processes won't necessarily know what's inside the buffers they receive, take a look at the \s-1TEMPLATE\s0 \s-1USAGE\s0 section.
\$1
Creates a new buffer object and returns it. The buffer is initially empty.
This method takes no arguments. Creates a new buffer object and appends to it each of the octet strings in @strs.
Returns the new buffer object. Returns the next 8-bit integer from the buffer (which is really just the \s-1ASCII\s0 code for the next character/byte in the buffer). Appends an 8-bit integer to the buffer (which is really just the character corresponding to that integer, in \s-1ASCII\s0). Returns the next 16-bit integer from the buffer. Appends a 16-bit integer to the buffer. Returns the next 32-bit integer from the buffer. Appends a 32-bit integer to the buffer. More appropriately called get_byte, perhaps, this returns the next byte from the buffer. Appends a byte (or a sequence of bytes) to the buffer. There is no restriction on the length of the byte string $bytes; if it makes you uncomfortable to call put_char to put multiple bytes, you can instead call this method as put_chars. It's the same thing. Grabs $n bytes from the buffer, where $n is a positive integer. Increments the internal offset state by $n. Appends a sequence of bytes to the buffer; if $n is unspecified, appends the entire length of $bytes. Otherwise appends only the first $n bytes of $bytes. Returns the next \*(L"string\*(R" from the buffer. A string here is represented as the length of the string (a 32-bit integer) followed by the string itself. Appends a string (32-bit integer length and the string itself) to the buffer. Extracts the next $n bytes from the buffer $buffer, increments the offset state in $buffer, and returns a new buffer object containing the extracted bytes.
Generally when you use Data::Buffer it's to communicate with another process (perhaps a C program) that bundles up its data into binary buffers. In those cases, it's very likely that the data will be in some well-known order in the buffer: in other words, it might be documented that a certain C program creates a buffer containing:
In this case, you would presumably know about the order of the data in the buffer, and you could extract it accordingly:
$buffer->get_int8; $buffer->get_str; $buffer->get_int32;
In other cases, however, there may not be a well-defined order of data items in the buffer. This might be the case if you're inventing your own protocol, and you want your binary buffers to \*(L"know\*(R" about their contents. In this case, you'll want to use the templating features of Data::Buffer.
When you use the put_ methods to place data in a buffer, Data::Buffer keeps track of the types of data that you're inserting in a template description of the buffer. This template contains all of the information necessary for a process to receive a buffer and extract the data in the buffer without knowledge of the order of the items.
To use this feature, simply use the insert_template method after you've filled your buffer to completion. For example:
my $buffer = Data::Buffer->new; $buffer->put_str("foo"); $buffer->put_int32(9999); $buffer->insert_template;
## Ship off the buffer to another process.
The receiving process should then invoke the get_all method on the buffer to extract all of the data:
my $buffer = Data::Buffer->new; $buffer->append( $received_buffer_data ); my @data = $buffer->get_all;
@data will now contain two elements: "foo" and 9999.
Appends raw data $bytes to the end of the in-memory buffer. Generally you don't need to use this method unless you're initializing an empty buffer, because when you need to add data to a buffer you should generally use one of the put_* methods. Empties out the buffer object. Behaves exactly like the substr built-in function, except on the buffer $buffer. Given no arguments, bytes returns the entire buffer; given one argument $offset, returns everything from that position to the end of the string; given $offset and $length, returns the segment of the buffer starting at $offset and consisting of $length bytes; and given all three arguments, replaces that segment with $replacement.
This is a very low-level method, and you generally won't need to use it.
Also be warned that you should not intermix use of this method with use of the get_* and put_* methods; the latter classes of methods maintain internal state of the buffer offset where arguments will be gotten from and put, respectively. The bytes method gives no thought to this internal offset state. Returns the length of the buffer object. Returns the internal offset state.
If you insist on intermixing calls to bytes with calls to the get_* and put_* methods, you'll probably want to use this method to get some status on that internal offset. Sets the internal offset state to $offset. Sets the internal offset state to 0. Returns a hex dump of the buffer. The dump is of the entire buffer $buffer; in other words, dump doesn't respect the internal offset pointer.
@args is passed directly through to the bytes method, which means that you can supply arguments to emulate support of the internal offset:
my $dump = $buffer->dump($buffer->offset); A helper method: pads out the buffer so that the length of the transferred packet will be evenly divisible by 8, which is a requirement of the \s-1SSH\s0 protocol.
Benjamin Trott, [email protected]
Except where otherwise noted, Data::Buffer is Copyright 2001 Benjamin Trott. All rights reserved. Data::Buffer is free software; you may redistribute it and/or modify it under the same terms as Perl itself.