== _test_ is surprisingly smart

[[Via Hacker News https://news.ycombinator.com/item?id=6751517]] I
would up reading [[Common shell script mistakes
http://www.pixelbeat.org/programming/shell_script_mistakes.html]]. When
I read this, I initially thought that it contained well-intentioned
but mistaken advice about _test_ (aka '_[ ... ]_'). Then I actually
checked what _test_'s behavior is and got a bunch of surprised. It
turns out that _test_ is really quite smart, sometimes disturbingly
so.

.pn prewrap on

Here's two different versions of a _test_ expression:

>  [ x"$var" = x"find" ] && echo yes
>  [ "$var" = "find" ] && echo yes

In theory, the reason the first version has an 'x' in front of both
sides is to deal with the case where someone sets _$var_ to something
that is a valid _test_ operator, like '_-a_' or '_-x_' or even
'_(_'; after all, '_[ -a = find ]_' doesn't look like a valid _test_
expression.  But if you actually check, it turns out that the second
version works perfectly well too.

What's going on is that _test_ is much smarter than you might think.
Rather than simply processing its arguments left to right, it uses
a much more complicated process of actually parsing its command
line.  When I started writing this entry I thought it was just modern
versions that behaved this way, but in fact the behavior is much older
than that; it goes all the way back to [[the V7 version of _test_
http://minnie.tuhs.org/cgi-bin/utree.pl?file=V7/usr/src/cmd/test.c]],
which actually implements a little recursive descent
parser (in quite readable code). This behavior is even
specified in [[the Single Unix Specification page for _test_
http://pubs.opengroup.org/onlinepubs/9699919799/utilities/test.html]]
where you can read the gory details for yourself (well, most of them).

(The exception is that the SuS version of _test_ doesn't include
_-a_ for and or _-o_ for or. This is an interesting exclusion since
it turns out they were actually in the V7 version of test [[per eg
the manpage
http://minnie.tuhs.org/cgi-bin/utree.pl?file=V7/usr/man/man1/test.1]].)

Note that this cleverness can break down in extreme situations. For
example, '_[ "$var1" -a "$var2" -a "$var3" ]_' is potentially
dangerous; consider what happens if _$var2_ is '_-r_'. And of course
you still really want to use _"..."_ to force things to be [[explicit
empty arguments EmptyArgumentVsNone]], because an outright missing
argument can easily completely change the meaning of a _test_
expression. Consider what happens to '_[ -r $var ]_' if _$var_ is
empty.

(It reduces to '_[ -r ]_', which is true because _-r_ is not the empty
string. You probably intended it to be false because a zero-length file
name is considered unreadable.)