Re^5: grep for hash?

It's not explicitly documented but it's stated in perlsyn that the assignment operator works "as in C", and it has the same right-to-left associativity as C's, and the value of the assignment is itself a valid rvalue for another subsequent assignment. Granted it's not chiseled into stone tablets, but I don't think that it's a "ZOMGWTFBBQ no one knows what'll happen just don't do that" case like ++$a + $a++ is.

That being said, your alternative with the separate assignment before use is definitely a clearer implementation.

The cake is a lie.
The cake is a lie.
The cake is a lie.

Comment on Re^5: grep for hash? Download Code

Replies are listed 'Best First'.
Re^6: grep for hash? by ikegami (Patriarch) on Jul 07, 2009 at 16:38 UTC
it has the same right-to-left associativity as C's Operator associativity is not related to operand evaluation order. Perl even have a case where they differ. Exponentiation is right-associative, but its operands are evaluated from left to right. it's stated in perlsyn that the assignment operator works "as in C" Operand evaluation order is not defined in C, and different compilers use different orders. (Well, so I was told.) but I don't think that it's a "ZOMGWTFBBQ no one knows what'll happen just don't do that I agree. Especially since `my $x = $x;` relies on it.	[reply] [d/l]
Re^7: grep for hash? by Fletch (Bishop) on Jul 07, 2009 at 16:54 UTC
That's one I didn't know; a new corner case to file away. `$ perl -le 'print +(do{ print STDERR "ORLY?"; 2})**(do{ print STDERR " +YARLY!"; 3})' ORLY? YARLY! 8` [download] The cake is a lie. The cake is a lie. The cake is a lie.	[reply] [d/l]