Re^3: Confused by Perl ternary operator

How is that execution-order unsafe? ?:'s first operand has to be evaluated before either the second or third.

After writing the above, I checked perlop.pod, and technically you are correct. ?: isn't documented as short-circuiting, so perl potentially could evaluate the 2nd and 3rd operands before testing the first one. However, I believe that it is intended to short-circuit and will submit a patch to the doc.

As an interesting side note, structures with else or elsif are implemented using condexpr (the ?: operator). That is,

if (foo) { bar } else { baz }
[download]

is identical to

foo ? do { bar } : do { baz }
[download]

so the short-circuiting of ?: is pretty guaranteed.

Comment on Re^3: Confused by Perl ternary operator Select or Download Code

Replies are listed 'Best First'.
Re^4: Confused by Perl ternary operator by diotalevi (Canon) on Aug 24, 2004 at 00:13 UTC
I was thinking that someone might use that expression that reads and writes to @_ inside another expression that depends on whether @_ has already been fetched or altered. ?: isn't just if/else, its an expression which can be used inside other expressions. `$join = ( @_ ? "Argument, the first" : shift ) . ( @_ ? "Argument, the second" : shift );` [download]	[reply] [d/l]