Interpolated code blocks may not create lexicals

[David Landgren <david@landgren.net>]

Porters,

I just managed to waste 10 minutes after having been bitten by the 
following parsing error:

print <<STR;
I like @{[
     my $rand = rand();
     $rand < 0.5 ? 'pie' : 'beer'
]}
STR

It is illegal, alas, to create a lexical in an interpolated ARRAY ref 
within a string. All is not lost since one can use an explicit do block 
to set up a pad:

print <<STR;
I like @{[
     do {
         my $rand = rand();
         $rand < 0.5 ? 'pie' : 'beer';
     }
]}
STR

I assume that setting up a block on the off chance that someone will do 
the above would incur a speed penalty for those cases where lexicals 
aren't created (which I would hope is the general case).

I had a look through the documentation (perldata, perlop, perlsyn, 
various perldoc -q Interpolation attempts) but didn't find anything. 
That's ok, I hardly ever do. Nor do I believe the documentation should 
be amended.

What I think should happen is that either my first attempt should Just 
Work, or else the error message

     syntax error at str line 2, near ");"
     syntax error at str line 4, near "'beer'
     ]"

should give a better hint, like 'cannot create a lexical in an 
interpolated code block'.

Beyond that, I have no idea how difficult this would be, and can offer 
little more than a TODO test to capture the desired semantics. So I had 
another look but didn't find a suitable test file in which to place it, 
so I made a new one, attached.

Thanks,
David
-- 
Stubborn tiny lights vs. clustering darkness forever ok?

interpstr.t

---

Follow-Ups from:

Nicholas Clark <nick@ccl4.org>
Aristotle Pagaltzis <pagaltzis@gmx.de>
Eirik Berg Hanssen <Eirik-Berg.Hanssen@allverden.no>

---

[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index][Thread Index][Top&Search][Original]

---