(Un)packing and (un)zipping in PHP and Ruby

Drab as a fool, aloof as a bard

2008-12-04 19:16

Posted by Hagfish in Programming

Comments (0)
Trackbacks (0)

(Un)packing and (un)zipping in PHP and Ruby

I like to think of myself as experienced in many computer languages, even if some of those experiences are unpleasant, and what better example of the variety of languages (and experiences) than PHP and Ruby. It seems that everyone skilled in one those languages has a strong preference for one or the other, so perhaps it would be helpful for me to contribute a comparison of how the two languages can be used to solve the same problem. My personal opinion is that while Ruby (and Rails sites) may fail spectacularly, PHP sites fail based on the quality of the programmer, and there are more PHP programmers out there (with a wider range of abilities) to pick bad examples of code from. Really, though, this blog post (and my whole blog) should not be about wading into the current argument of the day, and Ruby has its places and I’m happy to use it in those places, so I should make it clear that the comparison below is to aid interoperability. This was in fact what motivated me to write this code, actually, as there was a web service that needed implementing in both languages and the implementations were disagreeing on how the zipping stage should work. There are also a couple of gotchas just getting either language to zip or pack something, so hopefully the example code will be useful for someone trying to spot a bug in their own code.

PHP

<?php
$text = "abc";
echo "$text\n";

$zipped = gzcompress($text);
echo "$zipped\n";
echo chunk_split(bin2hex($zipped), 2, " ") . "\n";

$packed = pack("C*", 0x78, 0x9C, 0x4B, 0x4C, 0x4A, 0x06, 0x00, 0x02, 0x4D, 0x01, 0x27);
echo chunk_split(bin2hex($packed), 2, " ") . "\n";

$unzipped = gzuncompress($zipped);
echo "$unzipped\n";

$unzipped = gzuncompress($packed);
echo "$unzipped\n";
?>

The above code outputs:

abc
x�KLJM’
78 9c 4b 4c 4a 06 00 02 4d 01 27
78 9c 4b 4c 4a 06 00 02 4d 01 27
abc
abc

What is happening, if you want to follow line by line, is that PHP using the built-in gzcompress() function to compress the string abc, which becomes the (longer, because not every string can compress to something shorter) string x�KLJM’ but this is better represented as hex so bin2hex() is used to turn it into a string of hex digits. The same digits are then combined with pack() to create a binary string $packed which can be uncompressed with gzcompress() to the original abc again, the same as running gzcompress() on the $zipped string.

Ruby

require ‘zlib’

text = "abc"
puts text

zstream = Zlib::Deflate.new
zipped = zstream.deflate(text)
zipped << zstream.finish
zstream.close
puts zipped
puts zipped.unpack("C*").collect {|c| sprintf("%02X", c&0xff)}.join(’ ‘)

packed = [0x78, 0x9C, 0x4B, 0x4C, 0x4A, 0x06, 0x00, 0x02, 0x4D, 0x01, 0x27].pack(‘C*’)
puts packed.unpack("C*").collect {|c| sprintf("%02X", c&0xff)}.join(’ ‘)

zstream = Zlib::Inflate.new
unzipped = zstream.inflate(zipped)
zstream.close
puts unzipped

zstream = Zlib::Inflate.new
unzipped = zstream.inflate(packed)
zstream.close
puts unzipped

which produces an almost identical output:

abc
x�KLJM’
78 9C 4B 4C 4A 06 00 02 4D 01 27
78 9C 4B 4C 4A 06 00 02 4D 01 27
abc
abc

The things to notice are the use of included libraries (Zlib in this case) and accessing its functions using the :: syntax, the chaining of functions (like unpack(), collect() and join() on one line) without having to use nested brackets, and support for closures (using collect() with a free variable |c|). In terms of the trade off between using PHP and Ruby, notice that Ruby has annoying unguessable incantations like zipped << zstream.finish (only needed for the Deflate, not the Inflate, and occuring before the close), but that PHP has a confusing choice of Zlib functions.

Conclusion

It looks so simple, but the fact that other people on the net have different answers suggests there’s more to getting it right than meets the eye. Fortunately the rest of implementing the web service was easy in both languages and completely compatible, so to avoid blaming either of the languages we could just unfairly blame Zlib. Another deciding factor, of course, is support from hosting providers for the languages, where of course PHP wins, but with that attitude Groovy on Grails would never win.

If all software were written with loosely-coupled web service APIs, would we then have the freedom to code in whatever language we wanted?

(Un)packing and (un)zipping in PHP and Ruby

Drab as a fool, aloof as a bard

2008-12-04 19:16

(Un)packing and (un)zipping in PHP and Ruby

PHP

Ruby

Conclusion

Quicksearch

Archives

Categories

Syndicate This Blog

Blog Administration