Mac OS X Hacks

Dealing with Archives of Many Colors: .img, .sit, .tar, .gz

Back in the innocent days of OS 9, one compression format reigned supreme: Stuffit from Aladdin Systems. With OS X and its BSD Unix foundation, there’s a whole slew of compression technologies available, all built into your default installation.

Stuffit Expander, DropStuff, and their Aladdin ilk have long been stalwarts of the Mac OS, included on Apple CDs and preinstalled machines. The same can be said for Unix utilities like gzip, bzip2, and compress, also included with OS X and available through the Terminal. Throw in Apple’s disk-image technology, which creates archives that look and act like removable disks, and you’ve got a veritable cornucopia of compression and archival technologies.

.dmg and .img

Apple has been providing disk image technology in the shape of its Disk Copy utility for years now. Creating a disk image is a mindless task — simply open Disk Copy, drag a folder over the floating window (see Figure 1-11), decide if you want encryption, and choose where to save the resultant file (see Figure 1-12).

Figure 1-11. Dragging a folder into Disk Copy

Figure 1-12. Setting Image Folder options

Creating image files, however, doesn’t offer much compression, and you’ll see a lot of dmg.gz extensions on your new downloads. That leads us into gzip and tar. gzip is as much of a Unix standard as Stuffit has been for the Mac. By itself, it’s only a compression utility — it doesn’t bundle and archive multiple files like Aladdin’s DropStuff (also included in OS X). For that ability, it’s most often combined with another utility called tar or with the generated disk images from Apple’s Disk Copy. If you want to compress a .dmg file you’ve just created, you’d jump into the Terminal [Hack #48]:

gzip -9 filename.dmg

This command will automatically compress filename.dmg into filename.dmg.gz, at maximum compression. If we don’t include the -9, then gzip will finish slightly faster, but at the expense of a slightly larger file size (-6 is the default). Alternatively, if we’re going to use tar (very common when it comes to Unix downloads), we could bundle up our entire ~/Documents directory this way:

tar -cvf filename.tar ~/Documents

The c is to create a new archive, the v is to keep us informed of its progress, and f indicates the name of the final archive — in this case, filename.tar. Finally, we indicate what we want to archive, which is ~/Documents. We could easily archive more directories (or individual files) by adding them after our initial ~/Documents. Unlike gzip, tar only archives the files — it compresses nothing itself, much like Apple’s Disk Copy. To compress our new filename.tar, we’d used gzip as shown earlier. Because tar and gzip are so often intertwined, we can combine two commands into one:

tar -cvzf filename.tar.gz ~/Documents

Notice that we’ve added a z flag, which tells tar to automatically compress the final archive with gzip. We’ve also changed our final filename to reflect its compressed status. More information about both of these utilities can be accessed from your Terminal with man gzip and man tar.

bzip2

Whereas gzip uses a compression technique called Lempel-Ziv, bzip2 takes a different approach with the Burrows-Wheeler block-sorting text-compression algorithm. It’s a little slower compressing than gzip, but it often returns a smaller file size (see Table 1-1 at the end of this hack). Its use (and combination with tar) is similar to gzip, always preferring maximum compression:

tar -cvf filename.tar ~/Documents
bzip2 filename.tar

Other Compression Techniques

While gzip is more popular than bzip2 for Unix downloads, bzip2 has been making headway due to its stronger compression. Stuffit Expander can readily extract either format. Still more compression flavors exist, however. I’ve briefly outlined their usage here — you can find more information about their usage and specific abilities by typing man compress, man zip, or man jar in your Terminal.

# using the compress utility
tar -cvf filename.tar ~/Documents
compress filename.tar

# the same as previous
tar -cvZf filename.tar.Z ~/Documents

# now, zip at maximum compression
zip -r -9 filename.zip ~/Documents

# and jar (useful for Java applications)
jar cf filename.zip ~/Documents

Don’t Forget Stuffit

Aladdin Systems realized there would be a need for a simple drag and drop utility that could compress in other formats besides its own — that’s why you’ll see DropTar and DropZip utilities in your /Applications/Utilities/Stuffit Lite (or Stuffit Standard) directory. Using these is as you’d expect — simply drag and drop the files and folders you want to archive over its icon (or drag to its window), and you’re set. DropTar even has the capability to compress in multiple formats: bzip2, compress, gzip, and the native Stuffit format.

In Table 1-1, we’ve compressed a 100MB directory using each of the utilities, with maximum compression. If you’re looking for the smallest file, then bzip2 should be your first choice, but gzip could be more compatible with every computer your archive lands on (if you’re worried only about OS X, then bzip2 is a good bet). Be forewarned: the types of files you’re archiving will give you different results with each utility — the source directory in this case was filled with an equal amount of text, image, and binary files, but you’ll notice fluctuating results with large text files, multiple tiny files, and so on.

Table 1-1. Compression techniques and resulting file sizes

Compression technique	File size (in bytes)
compress	45,264,549
DropTar (compress)	45,032,503
jar	30,322,992
zip	30,232,529
DropTar (gzip)	30,069,414
gzip	30,042,941
DropZip	29,877,021
DropTar (bzip2)	26,072,415
bzip2	25,825,723