1PERLPACKTUT(1) Perl Programmers Reference Guide PERLPACKTUT(1)
2
6 perlpacktut - tutorial on "pack" and "unpack"
7
9 "pack" and "unpack" are two functions for transforming data according
10 to a user-defined template, between the guarded way Perl stores values
11 and some well-defined representation as might be required in the envi‐
12 ronment of a Perl program. Unfortunately, they're also two of the most
13 misunderstood and most often overlooked functions that Perl provides.
14 This tutorial will demystify them for you.
15
17 Most programming languages don't shelter the memory where variables are
18 stored. In C, for instance, you can take the address of some variable,
19 and the "sizeof" operator tells you how many bytes are allocated to the
20 variable. Using the address and the size, you may access the storage to
21 your heart's content.
22 In Perl, you just can't access memory at random, but the structural and
24 representational conversion provided by "pack" and "unpack" is an
25 excellent alternative. The "pack" function converts values to a byte
26 sequence containing representations according to a given specification,
27 the so-called "template" argument. "unpack" is the reverse process,
28 deriving some values from the contents of a string of bytes. (Be cau‐
29 tioned, however, that not all that has been packed together can be
30 neatly unpacked - a very common experience as seasoned travellers are
31 likely to confirm.)
32 Why, you may ask, would you need a chunk of memory containing some val‐
34 ues in binary representation? One good reason is input and output
35 accessing some file, a device, or a network connection, whereby this
36 binary representation is either forced on you or will give you some
37 benefit in processing. Another cause is passing data to some system
38 call that is not available as a Perl function: "syscall" requires you
39 to provide parameters stored in the way it happens in a C program. Even
40 text processing (as shown in the next section) may be simplified with
41 judicious usage of these two functions.
42 To see how (un)packing works, we'll start with a simple template code
44 where the conversion is in low gear: between the contents of a byte
45 sequence and a string of hexadecimal digits. Let's use "unpack", since
46 this is likely to remind you of a dump program, or some desperate last
47 message unfortunate programs are wont to throw at you before they
48 expire into the wild blue yonder. Assuming that the variable $mem holds
49 a sequence of bytes that we'd like to inspect without assuming anything
50 about its meaning, we can write
51 my( $hex ) = unpack( 'H*', $mem );
53 print "$hex\n";
54 whereupon we might see something like this, with each pair of hex dig‐
56 its corresponding to a byte:
57 41204d414e204120504c414e20412043414e414c2050414e414d41
59 What was in this chunk of memory? Numbers, characters, or a mixture of
61 both? Assuming that we're on a computer where ASCII (or some similar)
62 encoding is used: hexadecimal values in the range 0x40 - 0x5A indicate
63 an uppercase letter, and 0x20 encodes a space. So we might assume it is
64 a piece of text, which some are able to read like a tabloid; but others
65 will have to get hold of an ASCII table and relive that firstgrader
66 feeling. Not caring too much about which way to read this, we note that
67 "unpack" with the template code "H" converts the contents of a sequence
68 of bytes into the customary hexadecimal notation. Since "a sequence of"
69 is a pretty vague indication of quantity, "H" has been defined to con‐
70 vert just a single hexadecimal digit unless it is followed by a repeat
71 count. An asterisk for the repeat count means to use whatever remains.
72 The inverse operation - packing byte contents from a string of hexadec‐
74 imal digits - is just as easily written. For instance:
75 my $s = pack( 'H2' x 10, map { "3$_" } ( 0..9 ) );
77 print "$s\n";
78 Since we feed a list of ten 2-digit hexadecimal strings to "pack", the
80 pack template should contain ten pack codes. If this is run on a com‐
81 puter with ASCII character coding, it will print 0123456789.
82
84 Let's suppose you've got to read in a data file like this:
85 Date ⎪Description ⎪ Income⎪Expenditure
87 01/24/2001 Ahmed's Camel Emporium 1147.99
88 01/28/2001 Flea spray 24.99
89 01/29/2001 Camel rides to tourists 235.00
90 How do we do it? You might think first to use "split"; however, since
92 "split" collapses blank fields, you'll never know whether a record was
93 income or expenditure. Oops. Well, you could always use "substr":
94 while (<>) {
96 my $date = substr($_, 0, 11);
97 my $desc = substr($_, 12, 27);
98 my $income = substr($_, 40, 7);
99 my $expend = substr($_, 52, 7);
100 ...
101 }
102 It's not really a barrel of laughs, is it? In fact, it's worse than it
104 may seem; the eagle-eyed may notice that the first field should only be
105 10 characters wide, and the error has propagated right through the
106 other numbers - which we've had to count by hand. So it's error-prone
107 as well as horribly unfriendly.
108 Or maybe we could use regular expressions:
110 while (<>) {
112 my($date, $desc, $income, $expend) =
113 m⎪(\d\d/\d\d/\d{4}) (.{27}) (.{7})(.*)⎪;
114 ...
115 }
116 Urgh. Well, it's a bit better, but - well, would you want to maintain
118 that?
119 Hey, isn't Perl supposed to make this sort of thing easy? Well, it
121 does, if you use the right tools. "pack" and "unpack" are designed to
122 help you out when dealing with fixed-width data like the above. Let's
123 have a look at a solution with "unpack":
124 while (<>) {
126 my($date, $desc, $income, $expend) = unpack("A10xA27xA7A*", $_);
127 ...
128 }
129 That looks a bit nicer; but we've got to take apart that weird tem‐
131 plate. Where did I pull that out of?
132 OK, let's have a look at some of our data again; in fact, we'll include
134 the headers, and a handy ruler so we can keep track of where we are.
135 1 2 3 4 5
137 1234567890123456789012345678901234567890123456789012345678
138 Date ⎪Description ⎪ Income⎪Expenditure
139 01/28/2001 Flea spray 24.99
140 01/29/2001 Camel rides to tourists 235.00
141 From this, we can see that the date column stretches from column 1 to
143 column 10 - ten characters wide. The "pack"-ese for "character" is "A",
144 and ten of them are "A10". So if we just wanted to extract the dates,
145 we could say this:
146 my($date) = unpack("A10", $_);
148 OK, what's next? Between the date and the description is a blank col‐
150 umn; we want to skip over that. The "x" template means "skip forward",
151 so we want one of those. Next, we have another batch of characters,
152 from 12 to 38. That's 27 more characters, hence "A27". (Don't make the
153 fencepost error - there are 27 characters between 12 and 38, not 26.
154 Count 'em!)
155 Now we skip another character and pick up the next 7 characters:
157 my($date,$description,$income) = unpack("A10xA27xA7", $_);
159 Now comes the clever bit. Lines in our ledger which are just income and
161 not expenditure might end at column 46. Hence, we don't want to tell
162 our "unpack" pattern that we need to find another 12 characters; we'll
163 just say "if there's anything left, take it". As you might guess from
164 regular expressions, that's what the "*" means: "use everything remain‐
165 ing".
166 · Be warned, though, that unlike regular expressions, if the "unpack"
168 template doesn't match the incoming data, Perl will scream and die.
169 Hence, putting it all together:
171 my($date,$description,$income,$expend) = unpack("A10xA27xA7xA*", $_);
173 Now, that's our data parsed. I suppose what we might want to do now is
175 total up our income and expenditure, and add another line to the end of
176 our ledger - in the same format - saying how much we've brought in and
177 how much we've spent:
178 while (<>) {
180 my($date, $desc, $income, $expend) = unpack("A10xA27xA7xA*", $_);
181 $tot_income += $income;
182 $tot_expend += $expend;
183 }
184 $tot_income = sprintf("%.2f", $tot_income); # Get them into
186 $tot_expend = sprintf("%.2f", $tot_expend); # "financial" format
187 $date = POSIX::strftime("%m/%d/%Y", localtime);
189 # OK, let's go:
191 print pack("A10xA27xA7xA*", $date, "Totals", $tot_income, $tot_expend);
193 Oh, hmm. That didn't quite work. Let's see what happened:
195 01/24/2001 Ahmed's Camel Emporium 1147.99
197 01/28/2001 Flea spray 24.99
198 01/29/2001 Camel rides to tourists 1235.00
199 03/23/2001Totals 1235.001172.98
200 OK, it's a start, but what happened to the spaces? We put "x", didn't
202 we? Shouldn't it skip forward? Let's look at what "pack" in perlfunc
203 says:
204 x A null byte.
206 Urgh. No wonder. There's a big difference between "a null byte", char‐
208 acter zero, and "a space", character 32. Perl's put something between
209 the date and the description - but unfortunately, we can't see it!
210 What we actually need to do is expand the width of the fields. The "A"
212 format pads any non-existent characters with spaces, so we can use the
213 additional spaces to line up our fields, like this:
214 print pack("A11 A28 A8 A*", $date, "Totals", $tot_income, $tot_expend);
216 (Note that you can put spaces in the template to make it more readable,
218 but they don't translate to spaces in the output.) Here's what we got
219 this time:
220 01/24/2001 Ahmed's Camel Emporium 1147.99
222 01/28/2001 Flea spray 24.99
223 01/29/2001 Camel rides to tourists 1235.00
224 03/23/2001 Totals 1235.00 1172.98
225 That's a bit better, but we still have that last column which needs to
227 be moved further over. There's an easy way to fix this up: unfortu‐
228 nately, we can't get "pack" to right-justify our fields, but we can get
229 "sprintf" to do it:
230 $tot_income = sprintf("%.2f", $tot_income);
232 $tot_expend = sprintf("%12.2f", $tot_expend);
233 $date = POSIX::strftime("%m/%d/%Y", localtime);
234 print pack("A11 A28 A8 A*", $date, "Totals", $tot_income, $tot_expend);
235 This time we get the right answer:
237 01/28/2001 Flea spray 24.99
239 01/29/2001 Camel rides to tourists 1235.00
240 03/23/2001 Totals 1235.00 1172.98
241 So that's how we consume and produce fixed-width data. Let's recap what
243 we've seen of "pack" and "unpack" so far:
244 · Use "pack" to go from several pieces of data to one fixed-width ver‐
246 sion; use "unpack" to turn a fixed-width-format string into several
247 pieces of data.
248 · The pack format "A" means "any character"; if you're "pack"ing and
250 you've run out of things to pack, "pack" will fill the rest up with
251 spaces.
252 · "x" means "skip a byte" when "unpack"ing; when "pack"ing, it means
254 "introduce a null byte" - that's probably not what you mean if
255 you're dealing with plain text.
256 · You can follow the formats with numbers to say how many characters
258 should be affected by that format: "A12" means "take 12 characters";
259 "x6" means "skip 6 bytes" or "character 0, 6 times".
260 · Instead of a number, you can use "*" to mean "consume everything
262 else left".
263 Warning: when packing multiple pieces of data, "*" only means "con‐
265 sume all of the current piece of data". That's to say
266 pack("A*A*", $one, $two)
268 packs all of $one into the first "A*" and then all of $two into the
270 second. This is a general principle: each format character corre‐
271 sponds to one piece of data to be "pack"ed.
272
274 So much for textual data. Let's get onto the meaty stuff that "pack"
275 and "unpack" are best at: handling binary formats for numbers. There
276 is, of course, not just one binary format - life would be too simple -
277 but Perl will do all the finicky labor for you.
278 Integers
280 Packing and unpacking numbers implies conversion to and from some spe‐
282 cific binary representation. Leaving floating point numbers aside for
283 the moment, the salient properties of any such representation are:
284 · the number of bytes used for storing the integer,
286 · whether the contents are interpreted as a signed or unsigned num‐
288 ber,
289 · the byte ordering: whether the first byte is the least or most sig‐
291 nificant byte (or: little-endian or big-endian, respectively).
292 So, for instance, to pack 20302 to a signed 16 bit integer in your com‐
294 puter's representation you write
295 my $ps = pack( 's', 20302 );
297 Again, the result is a string, now containing 2 bytes. If you print
299 this string (which is, generally, not recommended) you might see "ON"
300 or "NO" (depending on your system's byte ordering) - or something
301 entirely different if your computer doesn't use ASCII character encod‐
302 ing. Unpacking $ps with the same template returns the original integer
303 value:
304 my( $s ) = unpack( 's', $ps );
306 This is true for all numeric template codes. But don't expect miracles:
308 if the packed value exceeds the allotted byte capacity, high order bits
309 are silently discarded, and unpack certainly won't be able to pull them
310 back out of some magic hat. And, when you pack using a signed template
311 code such as "s", an excess value may result in the sign bit getting
312 set, and unpacking this will smartly return a negative value.
313 16 bits won't get you too far with integers, but there is "l" and "L"
315 for signed and unsigned 32-bit integers. And if this is not enough and
316 your system supports 64 bit integers you can push the limits much
317 closer to infinity with pack codes "q" and "Q". A notable exception is
318 provided by pack codes "i" and "I" for signed and unsigned integers of
319 the "local custom" variety: Such an integer will take up as many bytes
320 as a local C compiler returns for "sizeof(int)", but it'll use at least
321 32 bits.
322 Each of the integer pack codes "sSlLqQ" results in a fixed number of
324 bytes, no matter where you execute your program. This may be useful for
325 some applications, but it does not provide for a portable way to pass
326 data structures between Perl and C programs (bound to happen when you
327 call XS extensions or the Perl function "syscall"), or when you read or
328 write binary files. What you'll need in this case are template codes
329 that depend on what your local C compiler compiles when you code
330 "short" or "unsigned long", for instance. These codes and their corre‐
331 sponding byte lengths are shown in the table below. Since the C stan‐
332 dard leaves much leeway with respect to the relative sizes of these
333 data types, actual values may vary, and that's why the values are given
334 as expressions in C and Perl. (If you'd like to use values from %Config
335 in your program you have to import it with "use Config".)
336 signed unsigned byte length in C byte length in Perl
338 s! S! sizeof(short) $Config{shortsize}
339 i! I! sizeof(int) $Config{intsize}
340 l! L! sizeof(long) $Config{longsize}
341 q! Q! sizeof(long long) $Config{longlongsize}
342 The "i!" and "I!" codes aren't different from "i" and "I"; they are
344 tolerated for completeness' sake.
345 Unpacking a Stack Frame
347 Requesting a particular byte ordering may be necessary when you work
349 with binary data coming from some specific architecture whereas your
350 program could run on a totally different system. As an example, assume
351 you have 24 bytes containing a stack frame as it happens on an Intel
352 8086:
353 +---------+ +----+----+ +---------+
355 TOS: ⎪ IP ⎪ TOS+4:⎪ FL ⎪ FH ⎪ FLAGS TOS+14:⎪ SI ⎪
356 +---------+ +----+----+ +---------+
357 ⎪ CS ⎪ ⎪ AL ⎪ AH ⎪ AX ⎪ DI ⎪
358 +---------+ +----+----+ +---------+
359 ⎪ BL ⎪ BH ⎪ BX ⎪ BP ⎪
360 +----+----+ +---------+
361 ⎪ CL ⎪ CH ⎪ CX ⎪ DS ⎪
362 +----+----+ +---------+
363 ⎪ DL ⎪ DH ⎪ DX ⎪ ES ⎪
364 +----+----+ +---------+
365 First, we note that this time-honored 16-bit CPU uses little-endian
367 order, and that's why the low order byte is stored at the lower
368 address. To unpack such a (signed) short we'll have to use code "v". A
369 repeat count unpacks all 12 shorts:
370 my( $ip, $cs, $flags, $ax, $bx, $cd, $dx, $si, $di, $bp, $ds, $es ) =
372 unpack( 'v12', $frame );
373 Alternatively, we could have used "C" to unpack the individually acces‐
375 sible byte registers FL, FH, AL, AH, etc.:
376 my( $fl, $fh, $al, $ah, $bl, $bh, $cl, $ch, $dl, $dh ) =
378 unpack( 'C10', substr( $frame, 4, 10 ) );
379 It would be nice if we could do this in one fell swoop: unpack a short,
381 back up a little, and then unpack 2 bytes. Since Perl is nice, it prof‐
382 fers the template code "X" to back up one byte. Putting this all
383 together, we may now write:
384 my( $ip, $cs,
386 $flags,$fl,$fh,
387 $ax,$al,$ah, $bx,$bl,$bh, $cx,$cl,$ch, $dx,$dl,$dh,
388 $si, $di, $bp, $ds, $es ) =
389 unpack( 'v2' . ('vXXCC' x 5) . 'v5', $frame );
390 (The clumsy construction of the template can be avoided - just read
392 on!)
393 We've taken some pains to construct the template so that it matches the
395 contents of our frame buffer. Otherwise we'd either get undefined val‐
396 ues, or "unpack" could not unpack all. If "pack" runs out of items, it
397 will supply null strings (which are coerced into zeroes whenever the
398 pack code says so).
399 How to Eat an Egg on a Net
401 The pack code for big-endian (high order byte at the lowest address) is
403 "n" for 16 bit and "N" for 32 bit integers. You use these codes if you
404 know that your data comes from a compliant architecture, but, surpris‐
405 ingly enough, you should also use these pack codes if you exchange
406 binary data, across the network, with some system that you know next to
407 nothing about. The simple reason is that this order has been chosen as
408 the network order, and all standard-fearing programs ought to follow
409 this convention. (This is, of course, a stern backing for one of the
410 Lilliputian parties and may well influence the political development
411 there.) So, if the protocol expects you to send a message by sending
412 the length first, followed by just so many bytes, you could write:
413 my $buf = pack( 'N', length( $msg ) ) . $msg;
415 or even:
417 my $buf = pack( 'NA*', length( $msg ), $msg );
419 and pass $buf to your send routine. Some protocols demand that the
421 count should include the length of the count itself: then just add 4 to
422 the data length. (But make sure to read "Lengths and Widths" before you
423 really code this!)
424 Floating point Numbers
426 For packing floating point numbers you have the choice between the pack
428 codes "f" and "d" which pack into (or unpack from) single-precision or
429 double-precision representation as it is provided by your system.
430 (There is no such thing as a network representation for reals, so if
431 you want to send your real numbers across computer boundaries, you'd
432 better stick to ASCII representation, unless you're absolutely sure
433 what's on the other end of the line.)
434
436 Bit Strings
437 Bits are the atoms in the memory world. Access to individual bits may
439 have to be used either as a last resort or because it is the most con‐
440 venient way to handle your data. Bit string (un)packing converts
441 between strings containing a series of 0 and 1 characters and a
442 sequence of bytes each containing a group of 8 bits. This is almost as
443 simple as it sounds, except that there are two ways the contents of a
444 byte may be written as a bit string. Let's have a look at an annotated
445 byte:
446 7 6 5 4 3 2 1 0
448 +-----------------+
449 ⎪ 1 0 0 0 1 1 0 0 ⎪
450 +-----------------+
451 MSB LSB
452 It's egg-eating all over again: Some think that as a bit string this
454 should be written "10001100" i.e. beginning with the most significant
455 bit, others insist on "00110001". Well, Perl isn't biased, so that's
456 why we have two bit string codes:
457 $byte = pack( 'B8', '10001100' ); # start with MSB
459 $byte = pack( 'b8', '00110001' ); # start with LSB
460 It is not possible to pack or unpack bit fields - just integral bytes.
462 "pack" always starts at the next byte boundary and "rounds up" to the
463 next multiple of 8 by adding zero bits as required. (If you do want bit
464 fields, there is "vec" in perlfunc. Or you could implement bit field
465 handling at the character string level, using split, substr, and con‐
466 catenation on unpacked bit strings.)
467 To illustrate unpacking for bit strings, we'll decompose a simple sta‐
469 tus register (a "-" stands for a "reserved" bit):
470 +-----------------+-----------------+
472 ⎪ S Z - A - P - C ⎪ - - - - O D I T ⎪
473 +-----------------+-----------------+
474 MSB LSB MSB LSB
475 Converting these two bytes to a string can be done with the unpack tem‐
477 plate 'b16'. To obtain the individual bit values from the bit string we
478 use "split" with the "empty" separator pattern which dissects into
479 individual characters. Bit values from the "reserved" positions are
480 simply assigned to "undef", a convenient notation for "I don't care
481 where this goes".
482 ($carry, undef, $parity, undef, $auxcarry, undef, $zero, $sign,
484 $trace, $interrupt, $direction, $overflow) =
485 split( //, unpack( 'b16', $status ) );
486 We could have used an unpack template 'b12' just as well, since the
488 last 4 bits can be ignored anyway.
489 Uuencoding
491 Another odd-man-out in the template alphabet is "u", which packs an
493 "uuencoded string". ("uu" is short for Unix-to-Unix.) Chances are that
494 you won't ever need this encoding technique which was invented to over‐
495 come the shortcomings of old-fashioned transmission mediums that do not
496 support other than simple ASCII data. The essential recipe is simple:
497 Take three bytes, or 24 bits. Split them into 4 six-packs, adding a
498 space (0x20) to each. Repeat until all of the data is blended. Fold
499 groups of 4 bytes into lines no longer than 60 and garnish them in
500 front with the original byte count (incremented by 0x20) and a "\n" at
501 the end. - The "pack" chef will prepare this for you, a la minute, when
502 you select pack code "u" on the menu:
503 my $uubuf = pack( 'u', $bindat );
505 A repeat count after "u" sets the number of bytes to put into an uuen‐
507 coded line, which is the maximum of 45 by default, but could be set to
508 some (smaller) integer multiple of three. "unpack" simply ignores the
509 repeat count.
510 Doing Sums
512 An even stranger template code is "%"<number>. First, because it's used
514 as a prefix to some other template code. Second, because it cannot be
515 used in "pack" at all, and third, in "unpack", doesn't return the data
516 as defined by the template code it precedes. Instead it'll give you an
517 integer of number bits that is computed from the data value by doing
518 sums. For numeric unpack codes, no big feat is achieved:
519 my $buf = pack( 'iii', 100, 20, 3 );
521 print unpack( '%32i3', $buf ), "\n"; # prints 123
522 For string values, "%" returns the sum of the byte values saving you
524 the trouble of a sum loop with "substr" and "ord":
525 print unpack( '%32A*', "\x01\x10" ), "\n"; # prints 17
527 Although the "%" code is documented as returning a "checksum": don't
529 put your trust in such values! Even when applied to a small number of
530 bytes, they won't guarantee a noticeable Hamming distance.
531 In connection with "b" or "B", "%" simply adds bits, and this can be
533 put to good use to count set bits efficiently:
534 my $bitcount = unpack( '%32b*', $mask );
536 And an even parity bit can be determined like this:
538 my $evenparity = unpack( '%1b*', $mask );
540 Unicode
542 Unicode is a character set that can represent most characters in most
544 of the world's languages, providing room for over one million different
545 characters. Unicode 3.1 specifies 94,140 characters: The Basic Latin
546 characters are assigned to the numbers 0 - 127. The Latin-1 Supplement
547 with characters that are used in several European languages is in the
548 next range, up to 255. After some more Latin extensions we find the
549 character sets from languages using non-Roman alphabets, interspersed
550 with a variety of symbol sets such as currency symbols, Zapf Dingbats
551 or Braille. (You might want to visit www.unicode.org for a look at
552 some of them - my personal favourites are Telugu and Kannada.)
553 The Unicode character sets associates characters with integers. Encod‐
555 ing these numbers in an equal number of bytes would more than double
556 the requirements for storing texts written in Latin alphabets. The
557 UTF-8 encoding avoids this by storing the most common (from a western
558 point of view) characters in a single byte while encoding the rarer
559 ones in three or more bytes.
560 So what has this got to do with "pack"? Well, if you want to convert
562 between a Unicode number and its UTF-8 representation you can do so by
563 using template code "U". As an example, let's produce the UTF-8 repre‐
564 sentation of the Euro currency symbol (code number 0x20AC):
565 $UTF8{Euro} = pack( 'U', 0x20AC );
567 Inspecting $UTF8{Euro} shows that it contains 3 bytes: "\xe2\x82\xac".
569 The round trip can be completed with "unpack":
570 $Unicode{Euro} = unpack( 'U', $UTF8{Euro} );
572 Usually you'll want to pack or unpack UTF-8 strings:
574 # pack and unpack the Hebrew alphabet
576 my $alefbet = pack( 'U*', 0x05d0..0x05ea );
577 my @hebrew = unpack( 'U*', $utf );
578 Another Portable Binary Encoding
580 The pack code "w" has been added to support a portable binary data
582 encoding scheme that goes way beyond simple integers. (Details can be
583 found at Casbah.org, the Scarab project.) A BER (Binary Encoded Repre‐
584 sentation) compressed unsigned integer stores base 128 digits, most
585 significant digit first, with as few digits as possible. Bit eight
586 (the high bit) is set on each byte except the last. There is no size
587 limit to BER encoding, but Perl won't go to extremes.
588 my $berbuf = pack( 'w*', 1, 128, 128+1, 128*128+127 );
590 A hex dump of $berbuf, with spaces inserted at the right places, shows
592 01 8100 8101 81807F. Since the last byte is always less than 128,
593 "unpack" knows where to stop.
594
596 Prior to Perl 5.8, repetitions of templates had to be made by "x"-mul‐
597 tiplication of template strings. Now there is a better way as we may
598 use the pack codes "(" and ")" combined with a repeat count. The
599 "unpack" template from the Stack Frame example can simply be written
600 like this:
601 unpack( 'v2 (vXXCC)5 v5', $frame )
603 Let's explore this feature a little more. We'll begin with the equiva‐
605 lent of
606 join( '', map( substr( $_, 0, 1 ), @str ) )
608 which returns a string consisting of the first character from each
610 string. Using pack, we can write
611 pack( '(A)'.@str, @str )
613 or, because a repeat count "*" means "repeat as often as required",
615 simply
616 pack( '(A)*', @str )
618 (Note that the template "A*" would only have packed $str[0] in full
620 length.)
621 To pack dates stored as triplets ( day, month, year ) in an array
623 @dates into a sequence of byte, byte, short integer we can write
624 $pd = pack( '(CCS)*', map( @$_, @dates ) );
626 To swap pairs of characters in a string (with even length) one could
628 use several techniques. First, let's use "x" and "X" to skip forward
629 and back:
630 $s = pack( '(A)*', unpack( '(xAXXAx)*', $s ) );
632 We can also use "@" to jump to an offset, with 0 being the position
634 where we were when the last "(" was encountered:
635 $s = pack( '(A)*', unpack( '(@1A @0A @2)*', $s ) );
637 Finally, there is also an entirely different approach by unpacking big
639 endian shorts and packing them in the reverse byte order:
640 $s = pack( '(v)*', unpack( '(n)*', $s );
642
644 String Lengths
645 In the previous section we've seen a network message that was con‐
647 structed by prefixing the binary message length to the actual message.
648 You'll find that packing a length followed by so many bytes of data is
649 a frequently used recipe since appending a null byte won't work if a
650 null byte may be part of the data. Here is an example where both tech‐
651 niques are used: after two null terminated strings with source and des‐
652 tination address, a Short Message (to a mobile phone) is sent after a
653 length byte:
654 my $msg = pack( 'Z*Z*CA*', $src, $dst, length( $sm ), $sm );
656 Unpacking this message can be done with the same template:
658 ( $src, $dst, $len, $sm ) = unpack( 'Z*Z*CA*', $msg );
660 There's a subtle trap lurking in the offing: Adding another field after
662 the Short Message (in variable $sm) is all right when packing, but this
663 cannot be unpacked naively:
664 # pack a message
666 my $msg = pack( 'Z*Z*CA*C', $src, $dst, length( $sm ), $sm, $prio );
667 # unpack fails - $prio remains undefined!
669 ( $src, $dst, $len, $sm, $prio ) = unpack( 'Z*Z*CA*C', $msg );
670 The pack code "A*" gobbles up all remaining bytes, and $prio remains
672 undefined! Before we let disappointment dampen the morale: Perl's got
673 the trump card to make this trick too, just a little further up the
674 sleeve. Watch this:
675 # pack a message: ASCIIZ, ASCIIZ, length/string, byte
677 my $msg = pack( 'Z* Z* C/A* C', $src, $dst, $sm, $prio );
678 # unpack
680 ( $src, $dst, $sm, $prio ) = unpack( 'Z* Z* C/A* C', $msg );
681 Combining two pack codes with a slash ("/") associates them with a sin‐
683 gle value from the argument list. In "pack", the length of the argument
684 is taken and packed according to the first code while the argument
685 itself is added after being converted with the template code after the
686 slash. This saves us the trouble of inserting the "length" call, but
687 it is in "unpack" where we really score: The value of the length byte
688 marks the end of the string to be taken from the buffer. Since this
689 combination doesn't make sense except when the second pack code isn't
690 "a*", "A*" or "Z*", Perl won't let you.
691 The pack code preceding "/" may be anything that's fit to represent a
693 number: All the numeric binary pack codes, and even text codes such as
694 "A4" or "Z*":
695 # pack/unpack a string preceded by its length in ASCII
697 my $buf = pack( 'A4/A*', "Humpty-Dumpty" );
698 # unpack $buf: '13 Humpty-Dumpty'
699 my $txt = unpack( 'A4/A*', $buf );
700 "/" is not implemented in Perls before 5.6, so if your code is required
702 to work on older Perls you'll need to "unpack( 'Z* Z* C')" to get the
703 length, then use it to make a new unpack string. For example
704 # pack a message: ASCIIZ, ASCIIZ, length, string, byte (5.005 compatible)
706 my $msg = pack( 'Z* Z* C A* C', $src, $dst, length $sm, $sm, $prio );
707 # unpack
709 ( undef, undef, $len) = unpack( 'Z* Z* C', $msg );
710 ($src, $dst, $sm, $prio) = unpack ( "Z* Z* x A$len C", $msg );
711 But that second "unpack" is rushing ahead. It isn't using a simple lit‐
713 eral string for the template. So maybe we should introduce...
714 Dynamic Templates
716 So far, we've seen literals used as templates. If the list of pack
718 items doesn't have fixed length, an expression constructing the tem‐
719 plate is required (whenever, for some reason, "()*" cannot be used).
720 Here's an example: To store named string values in a way that can be
721 conveniently parsed by a C program, we create a sequence of names and
722 null terminated ASCII strings, with "=" between the name and the value,
723 followed by an additional delimiting null byte. Here's how:
724 my $env = pack( '(A*A*Z*)' . keys( %Env ) . 'C',
726 map( { ( $_, '=', $Env{$_} ) } keys( %Env ) ), 0 );
727 Let's examine the cogs of this byte mill, one by one. There's the "map"
729 call, creating the items we intend to stuff into the $env buffer: to
730 each key (in $_) it adds the "=" separator and the hash entry value.
731 Each triplet is packed with the template code sequence "A*A*Z*" that is
732 repeated according to the number of keys. (Yes, that's what the "keys"
733 function returns in scalar context.) To get the very last null byte, we
734 add a 0 at the end of the "pack" list, to be packed with "C". (Atten‐
735 tive readers may have noticed that we could have omitted the 0.)
736 For the reverse operation, we'll have to determine the number of items
738 in the buffer before we can let "unpack" rip it apart:
739 my $n = $env =~ tr/\0// - 1;
741 my %env = map( split( /=/, $_ ), unpack( "(Z*)$n", $env ) );
742 The "tr" counts the null bytes. The "unpack" call returns a list of
744 name-value pairs each of which is taken apart in the "map" block.
745 Counting Repetitions
747 Rather than storing a sentinel at the end of a data item (or a list of
749 items), we could precede the data with a count. Again, we pack keys and
750 values of a hash, preceding each with an unsigned short length count,
751 and up front we store the number of pairs:
752 my $env = pack( 'S(S/A* S/A*)*', scalar keys( %Env ), %Env );
754 This simplifies the reverse operation as the number of repetitions can
756 be unpacked with the "/" code:
757 my %env = unpack( 'S/(S/A* S/A*)', $env );
759 Note that this is one of the rare cases where you cannot use the same
761 template for "pack" and "unpack" because "pack" can't determine a
762 repeat count for a "()"-group.
763
765 In previous sections we have seen how to pack numbers and character
766 strings. If it were not for a couple of snags we could conclude this
767 section right away with the terse remark that C structures don't con‐
768 tain anything else, and therefore you already know all there is to it.
769 Sorry, no: read on, please.
770 The Alignment Pit
772 In the consideration of speed against memory requirements the balance
774 has been tilted in favor of faster execution. This has influenced the
775 way C compilers allocate memory for structures: On architectures where
776 a 16-bit or 32-bit operand can be moved faster between places in mem‐
777 ory, or to or from a CPU register, if it is aligned at an even or mul‐
778 tiple-of-four or even at a multiple-of eight address, a C compiler will
779 give you this speed benefit by stuffing extra bytes into structures.
780 If you don't cross the C shoreline this is not likely to cause you any
781 grief (although you should care when you design large data structures,
782 or you want your code to be portable between architectures (you do want
783 that, don't you?)).
784 To see how this affects "pack" and "unpack", we'll compare these two C
786 structures:
787 typedef struct {
789 char c1;
790 short s;
791 char c2;
792 long l;
793 } gappy_t;
794 typedef struct {
796 long l;
797 short s;
798 char c1;
799 char c2;
800 } dense_t;
801 Typically, a C compiler allocates 12 bytes to a "gappy_t" variable, but
803 requires only 8 bytes for a "dense_t". After investigating this fur‐
804 ther, we can draw memory maps, showing where the extra 4 bytes are hid‐
805 den:
806 0 +4 +8 +12
808 +--+--+--+--+--+--+--+--+--+--+--+--+
809 ⎪c1⎪xx⎪ s ⎪c2⎪xx⎪xx⎪xx⎪ l ⎪ xx = fill byte
810 +--+--+--+--+--+--+--+--+--+--+--+--+
811 gappy_t
812 0 +4 +8
814 +--+--+--+--+--+--+--+--+
815 ⎪ l ⎪ h ⎪c1⎪c2⎪
816 +--+--+--+--+--+--+--+--+
817 dense_t
818 And that's where the first quirk strikes: "pack" and "unpack" templates
820 have to be stuffed with "x" codes to get those extra fill bytes.
821 The natural question: "Why can't Perl compensate for the gaps?" war‐
823 rants an answer. One good reason is that C compilers might provide
824 (non-ANSI) extensions permitting all sorts of fancy control over the
825 way structures are aligned, even at the level of an individual struc‐
826 ture field. And, if this were not enough, there is an insidious thing
827 called "union" where the amount of fill bytes cannot be derived from
828 the alignment of the next item alone.
829 OK, so let's bite the bullet. Here's one way to get the alignment right
831 by inserting template codes "x", which don't take a corresponding item
832 from the list:
833 my $gappy = pack( 'cxs cxxx l!', $c1, $s, $c2, $l );
835 Note the "!" after "l": We want to make sure that we pack a long inte‐
837 ger as it is compiled by our C compiler. And even now, it will only
838 work for the platforms where the compiler aligns things as above. And
839 somebody somewhere has a platform where it doesn't. [Probably a Cray,
840 where "short"s, "int"s and "long"s are all 8 bytes. :-)]
841 Counting bytes and watching alignments in lengthy structures is bound
843 to be a drag. Isn't there a way we can create the template with a sim‐
844 ple program? Here's a C program that does the trick:
845 #include <stdio.h>
847 #include <stddef.h>
848 typedef struct {
850 char fc1;
851 short fs;
852 char fc2;
853 long fl;
854 } gappy_t;
855 #define Pt(struct,field,tchar) \
857 printf( "@%d%s ", offsetof(struct,field), # tchar );
858 int main() {
860 Pt( gappy_t, fc1, c );
861 Pt( gappy_t, fs, s! );
862 Pt( gappy_t, fc2, c );
863 Pt( gappy_t, fl, l! );
864 printf( "\n" );
865 }
866 The output line can be used as a template in a "pack" or "unpack" call:
868 my $gappy = pack( '@0c @2s! @4c @8l!', $c1, $s, $c2, $l );
870 Gee, yet another template code - as if we hadn't plenty. But "@" saves
872 our day by enabling us to specify the offset from the beginning of the
873 pack buffer to the next item: This is just the value the "offsetof"
874 macro (defined in "<stddef.h>") returns when given a "struct" type and
875 one of its field names ("member-designator" in C standardese).
876 Neither using offsets nor adding "x"'s to bridge the gaps is satisfac‐
878 tory. (Just imagine what happens if the structure changes.) What we
879 really need is a way of saying "skip as many bytes as required to the
880 next multiple of N". In fluent Templatese, you say this with "x!N"
881 where N is replaced by the appropriate value. Here's the next version
882 of our struct packaging:
883 my $gappy = pack( 'c x!2 s c x!4 l!', $c1, $s, $c2, $l );
885 That's certainly better, but we still have to know how long all the
887 integers are, and portability is far away. Rather than 2, for instance,
888 we want to say "however long a short is". But this can be done by
889 enclosing the appropriate pack code in brackets: "[s]". So, here's the
890 very best we can do:
891 my $gappy = pack( 'c x![s] s c x![l!] l!', $c1, $s, $c2, $l );
893 Alignment, Take 2
895 I'm afraid that we're not quite through with the alignment catch yet.
897 The hydra raises another ugly head when you pack arrays of structures:
898 typedef struct {
900 short count;
901 char glyph;
902 } cell_t;
903 typedef cell_t buffer_t[BUFLEN];
905 Where's the catch? Padding is neither required before the first field
907 "count", nor between this and the next field "glyph", so why can't we
908 simply pack like this:
909 # something goes wrong here:
911 pack( 's!a' x @buffer,
912 map{ ( $_->{count}, $_->{glyph} ) } @buffer );
913 This packs "3*@buffer" bytes, but it turns out that the size of "buf‐
915 fer_t" is four times "BUFLEN"! The moral of the story is that the
916 required alignment of a structure or array is propagated to the next
917 higher level where we have to consider padding at the end of each com‐
918 ponent as well. Thus the correct template is:
919 pack( 's!ax' x @buffer,
921 map{ ( $_->{count}, $_->{glyph} ) } @buffer );
922 Alignment, Take 3
924 And even if you take all the above into account, ANSI still lets this:
926 typedef struct {
928 char foo[2];
929 } foo_t;
930 vary in size. The alignment constraint of the structure can be greater
932 than any of its elements. [And if you think that this doesn't affect
933 anything common, dismember the next cellphone that you see. Many have
934 ARM cores, and the ARM structure rules make "sizeof (foo_t)" == 4]
935 Pointers for How to Use Them
937 The title of this section indicates the second problem you may run into
939 sooner or later when you pack C structures. If the function you intend
940 to call expects a, say, "void *" value, you cannot simply take a refer‐
941 ence to a Perl variable. (Although that value certainly is a memory
942 address, it's not the address where the variable's contents are
943 stored.)
944 Template code "P" promises to pack a "pointer to a fixed length
946 string". Isn't this what we want? Let's try:
947 # allocate some storage and pack a pointer to it
949 my $memory = "\x00" x $size;
950 my $memptr = pack( 'P', $memory );
951 But wait: doesn't "pack" just return a sequence of bytes? How can we
953 pass this string of bytes to some C code expecting a pointer which is,
954 after all, nothing but a number? The answer is simple: We have to
955 obtain the numeric address from the bytes returned by "pack".
956 my $ptr = unpack( 'L!', $memptr );
958 Obviously this assumes that it is possible to typecast a pointer to an
960 unsigned long and vice versa, which frequently works but should not be
961 taken as a universal law. - Now that we have this pointer the next
962 question is: How can we put it to good use? We need a call to some C
963 function where a pointer is expected. The read(2) system call comes to
964 mind:
965 ssize_t read(int fd, void *buf, size_t count);
967 After reading perlfunc explaining how to use "syscall" we can write
969 this Perl function copying a file to standard output:
970 require 'syscall.ph';
972 sub cat($){
973 my $path = shift();
974 my $size = -s $path;
975 my $memory = "\x00" x $size; # allocate some memory
976 my $ptr = unpack( 'L', pack( 'P', $memory ) );
977 open( F, $path ) ⎪⎪ die( "$path: cannot open ($!)\n" );
978 my $fd = fileno(F);
979 my $res = syscall( &SYS_read, fileno(F), $ptr, $size );
980 print $memory;
981 close( F );
982 }
983 This is neither a specimen of simplicity nor a paragon of portability
985 but it illustrates the point: We are able to sneak behind the scenes
986 and access Perl's otherwise well-guarded memory! (Important note:
987 Perl's "syscall" does not require you to construct pointers in this
988 roundabout way. You simply pass a string variable, and Perl forwards
989 the address.)
990 How does "unpack" with "P" work? Imagine some pointer in the buffer
992 about to be unpacked: If it isn't the null pointer (which will smartly
993 produce the "undef" value) we have a start address - but then what?
994 Perl has no way of knowing how long this "fixed length string" is, so
995 it's up to you to specify the actual size as an explicit length after
996 "P".
997 my $mem = "abcdefghijklmn";
999 print unpack( 'P5', pack( 'P', $mem ) ); # prints "abcde"
1000 As a consequence, "pack" ignores any number or "*" after "P".
1002 Now that we have seen "P" at work, we might as well give "p" a whirl.
1004 Why do we need a second template code for packing pointers at all? The
1005 answer lies behind the simple fact that an "unpack" with "p" promises a
1006 null-terminated string starting at the address taken from the buffer,
1007 and that implies a length for the data item to be returned:
1008 my $buf = pack( 'p', "abc\x00efhijklmn" );
1010 print unpack( 'p', $buf ); # prints "abc"
1011 Albeit this is apt to be confusing: As a consequence of the length
1013 being implied by the string's length, a number after pack code "p" is a
1014 repeat count, not a length as after "P".
1015 Using "pack(..., $x)" with "P" or "p" to get the address where $x is
1017 actually stored must be used with circumspection. Perl's internal
1018 machinery considers the relation between a variable and that address as
1019 its very own private matter and doesn't really care that we have
1020 obtained a copy. Therefore:
1021 · Do not use "pack" with "p" or "P" to obtain the address of variable
1023 that's bound to go out of scope (and thereby freeing its memory)
1024 before you are done with using the memory at that address.
1025 · Be very careful with Perl operations that change the value of the
1027 variable. Appending something to the variable, for instance, might
1028 require reallocation of its storage, leaving you with a pointer
1029 into no-man's land.
1030 · Don't think that you can get the address of a Perl variable when it
1032 is stored as an integer or double number! "pack('P', $x)" will
1033 force the variable's internal representation to string, just as if
1034 you had written something like "$x .= ''".
1035 It's safe, however, to P- or p-pack a string literal, because Perl sim‐
1037 ply allocates an anonymous variable.
1038
1040 Here are a collection of (possibly) useful canned recipes for "pack"
1041 and "unpack":
1042 # Convert IP address for socket functions
1044 pack( "C4", split /\./, "123.4.5.6" );
1045 # Count the bits in a chunk of memory (e.g. a select vector)
1047 unpack( '%32b*', $mask );
1048 # Determine the endianness of your system
1050 $is_little_endian = unpack( 'c', pack( 's', 1 ) );
1051 $is_big_endian = unpack( 'xc', pack( 's', 1 ) );
1052 # Determine the number of bits in a native integer
1054 $bits = unpack( '%32I!', ~0 );
1055 # Prepare argument for the nanosleep system call
1057 my $timespec = pack( 'L!L!', $secs, $nanosecs );
1058 For a simple memory dump we unpack some bytes into just as many pairs
1060 of hex digits, and use "map" to handle the traditional spacing - 16
1061 bytes to a line:
1062 my $i;
1064 print map( ++$i % 16 ? "$_ " : "$_\n",
1065 unpack( 'H2' x length( $mem ), $mem ) ),
1066 length( $mem ) % 16 ? "\n" : '';
1067
1069 # Pulling digits out of nowhere...
1070 print unpack( 'C', pack( 'x' ) ),
1071 unpack( '%B*', pack( 'A' ) ),
1072 unpack( 'H', pack( 'A' ) ),
1073 unpack( 'A', unpack( 'C', pack( 'A' ) ) ), "\n";
1074 # One for the road ;-)
1076 my $advice = pack( 'all u can in a van' );
1077
1079 Simon Cozens and Wolfgang Laun.
1080perl v5.8.8 2006-01-07 PERLPACKTUT(1)