my $state = 1;
    state = 2;
[download]

-- Roman

Unfortunately, this is at $WORK, so I can’t easily try other versions.

In my code, the error-message was reported on a line ... other than the my $state; declaration ... where the offending variable did appear, just not the place where (it did!) occur as a bareword.   Hence, the code looked more like this:

  my $state = 0;
  ...
  $state = 1;     #<--- BAREWORD REPORTED HERE
  ...
  $state = 4;
  ...
  state = 5;      #<--- BUT THE BAREWORD IS ACTUALLY HERE
[download]

The message line-number was not simply “off by some random amount.”   The variable did indeed occur as a bareword, elsewhere in the program.   The message simply cited the wrong line.

I can not repeat your problem on my system with 5.10.0. The message points to the correct line:

$ cat foo.pl
my $state = 0;
print "state = $state\n";
$state = 1;
print "state = $state\n";
$state = 4;
print "state = $state\n";
state = 5;
print "state = $state\n";
$
$ perl -Mstrict foo.pl
Can't modify constant item in scalar assignment at foo.pl line 7, near
+ "5;"
Bareword "state" not allowed while "strict subs" in use at foo.pl line
+ 7.
Execution of foo.pl aborted due to compilation errors.
$
$ perl -v

This is perl, v5.10.0 built for x86_64-linux

Copyright 1987-2007, Larry Wall

Perl may be copied only under the terms of either the Artistic License
+ or the
GNU General Public License, which may be found in the Perl 5 source ki
+t.

Complete documentation for Perl, including FAQ lists, should be found 
+on
this system using "man perl" or "perldoc perl".  If you have access to
+ the
Internet, point your browser at http://www.perl.org/, the Perl Home Pa
+ge.

$
[download]

I also tried Perl versions 5.8.8 and 5.12.0.

That's an interesting theory. But, as you found, the effect is quite unfortunate for the one trying to fix the error. I would personally love to see an example that can reproduce the problem. And such would make it more likely that the bug gets fixed. Thanks.

- tye        

As I was fiddling with it, preparing it to copy here ... I couldn’t reproduce the error anymore.   This time, when I fiddled with the line, removing the "$", I got this message that I have not seen before:
Can't modify constant item in scalar assignment at alljobs_file.pl line 67, near "4;"

I have no idea why Perl would see it as “a constant item.”   (There is probably a clue lurking here, but I do not know what it would be.)

Very strange.   So, the behavior that I am seeing is now inconsistent.   (And yes, it is running against a “local Perl.”)   Anyhow, the code is here:

Don’t go looking for TWS::Parser in CPAN, because it doesn’t exist.   It is a parser that I am right now working-on to process Tivoli Workload Scheduler (e-e-e-e-e-e-yuck!) schedule and job files.   (And, by the way, it works.   For our very old version of TWS, anyway.   Anybody out there want such a thing?)

I am not inclined to say that this issue is “high priority” or anything.   It is not.   Rather, I wanted to document the behavior because it caught me by surprise and wasted me some time.   There is, as you can see, no switch demagoguery or other such nonsense going on here:   it’s just straight Perl.

(Blink...)

I just encountered a comment that is talking about “the state feature of Perl 5.10.”

I know nothing about this feature (yet...), but could it be that this has something to do with what I am seeing here?

(I’m not annoyed, nor show-stopped ... I’m curious now!)

Hi sundialsvc4,

It's a very cool feature of 5.10 which mimics the static variable of C.  State variables don't lose their data, even when they go out of scope, so you can use them for cases where you want data to persist, without having to use globals.

For example, to perform initialization only once the first time a subroutine is called, you could do this:

#!/usr/bin/perl -w

use feature qw{ state };
use strict;
use warnings;

for (1..5) {
    some_sub();
}

sub some_sub {
    (state $b_initialized++) or initialize();
    print "Now doing the main work of some_sub() ...\n";
}


sub initialize {
    print "Initializing....\n";
    # ... etc ...
}

__END__

=== Output ===

Initializing....
Now doing the main work of some_sub() ...
Now doing the main work of some_sub() ...
Now doing the main work of some_sub() ...
Now doing the main work of some_sub() ...
Now doing the main work of some_sub() ...
[download]

Or to define a subroutine which returns a unique integer each time it's called:

#!/usr/bin/perl -w

use feature qw{ state };
use strict;
use warnings;

for (1..5) {
    printf "Got the value %d\n", unique_integer();
}


sub unique_integer {
    return ++(state $ncalls);
}

__END__

=== Output ===

Got the value 1
Got the value 2
Got the value 3
Got the value 4
Got the value 5
[download]

---

s''(q.S:$/9=(T1';s;(..)(..);$..=substr+crypt($1,$2),2,3;eg;print$..$/

, but could it be that this has something to do with what I am seeing here?

No, absolutely not.