Order of execution of functions in list

vsespb has asked for the wisdom of the Perl Monks concerning the following question:

Replies are listed 'Best First'.
Re: Order of execution of functions in list by ikegami (Patriarch) on Sep 13, 2013 at 13:23 UTC
Operand evaluation order isn't defined for all operators^[1], but it is for this one. Quote perlop, In list context, [`,`] is just the list argument separator, and inserts both its arguments into the list. These arguments are also evaluated from left to right. That's perfectly fine. For example, Perl doesn't define the order in which `f`, `g` and `h` will be called by `f() + g() + h()`.	[reply] [d/l] [select]
Re^2: Order of execution of functions in list by LanX (Saint) on Sep 13, 2013 at 14:17 UTC
So in function calls it's not a list argument separator ? `function($i++,++$i)` ??? Cheers Rolf ( addicted to the Perl Programming Language)	[reply] [d/l]
Re^3: Order of execution of functions in list by ikegami (Patriarch) on Sep 16, 2013 at 02:02 UTC
So in function calls it's not a list argument separator ? Sure it is. `function($i++,++$i) ???` `$i++` is going to be evaluated before `++$i`.	[reply] [d/l] [select]
Re^2: Order of execution of functions in list by vsespb (Chaplain) on Sep 13, 2013 at 13:26 UTC
Ok, thanks! That was exactly what I looked for!	[reply]
Re: Order of execution of functions in list by ikegami (Patriarch) on Sep 13, 2013 at 13:59 UTC
Also, `my ($a1, $a2) = splice(@_, 0, 2);` [download]	[reply] [d/l]
Re: Order of execution of functions in list by LanX (Saint) on Sep 13, 2013 at 12:51 UTC
Reading the linked discussions reveals that the problem is related to lvalues and composed expressions (like `++$i`) with side effects. `shift` is a simple expression (a builtin in this case), has no side-effects and returns an rvalue. Cheers Rolf ( addicted to the Perl Programming Language)	[reply] [d/l] [select]
Re^2: Order of execution of functions in list by ikegami (Patriarch) on Sep 13, 2013 at 13:26 UTC
has no side-effects `$ perl -E'@a=qw( a b c ); say 0+@a; shift(@a); say 0+@a;' 3 2` [download] returns an rvalue `$ perl -E'$_=123; say; sub { sub { ++$_[0] }->(shift); }->($_); say;' 123 124` [download]	[reply] [d/l] [select]
Re^3: Order of execution of functions in list by LanX (Saint) on Sep 13, 2013 at 14:05 UTC
The problem with mobiles is that I can't check my code (at least ATM) And I don't understand the sideeffect code you posted. I'll reply this evening... Edit At second glance it seems that shift really returns lvalues... Strange Mea culpa!!! :( Cheers Rolf ( addicted to the Perl Programming Language)	[reply]
Re^3: Order of execution of functions in list by LanX (Saint) on Sep 13, 2013 at 17:02 UTC
> > returns an rvalue FWIW: I think this code is easier to understand and demonstrates your point `DB<194> sub tst { my $x=\(shift()); ++$$x } DB<195> $a=10 => 10 DB<196> tst $a => 11 DB<197> tst $a => 12 DB<198> $a => 12` [download] I get the same results using the equivalent `splice(@_,0,1)` That's not documented in the perldocs for shift or splice and I think it's due to the way Perl holds and replaces variables in @_, they just don't loose their aliasing magic when shifted. Maybe Perl avoids copying of the values for performance reasons. Interesting... But since it's not documented, I doubt that this is reliable behavior. Cheers Rolf ( addicted to the Perl Programming Language)	[reply] [d/l] [select]
Re^4: Order of execution of functions in list by ikegami (Patriarch) on Sep 16, 2013 at 02:08 UTC
Re^5: Order of execution of functions in list ( \(shift @_ ) reference to alias \$_[0] references original ) by Anonymous Monk on Sep 16, 2013 at 03:12 UTC
Some notes below your chosen depth have not been shown here
Re^2: Order of execution of functions in list by vsespb (Chaplain) on Sep 13, 2013 at 13:06 UTC
has no side-effects shift() actually modifies @_, isn't it side-effect? if shift()s executed in different order, result will be different.	[reply]
Re^3: Order of execution of functions in list by LanX (Saint) on Sep 13, 2013 at 14:00 UTC
Imho a main effect cause its a builtin function. Do you expect `sub inc { return $_[0] + 1}` to cause ambiguity ? Ok ... Maybe better phrased "a well defined effect" my definition of side was to restricted. Cheers Rolf ( addicted to the Perl Programming Language)	[reply] [d/l]
Re: Order of execution of functions in list by Marshall (Canon) on Sep 13, 2013 at 13:20 UTC
Well first, there is a problem with $a and $b in Perl. These are special variables used in for example the sort function. Use $x and $y instead. I have bench-marked the shift vs array list assignment. my $X = shift; my $Y = shift; vs my ($X,$Y) = @_; The second version is faster with more than one variable. `my ($a1, $a2) = (shift, shift);` [download] Better is: `my ($X,$Y) = @_;` [download]	[reply] [d/l] [select]
Re^2: Order of execution of functions in list by vsespb (Chaplain) on Sep 16, 2013 at 08:44 UTC
Well first, there is a problem with $a and $b in Perl 1. there is no $a and $b in my code. 2. it's not a problem, at least if you are not defining comparison function in same scope.	[reply]
Re^2: Order of execution of functions in list by vsespb (Chaplain) on Sep 13, 2013 at 13:29 UTC
That was just proof-of-concept example, my original code a bit more complex and I care more about readability than performance in my particular case.	[reply]

Edit