RewriteEngine on
RewriteRule /forums/viewtopic.php?f=12 http://www.newdomain.com/forums
+/

After a relatively prolonged absence (from cluttering this forum with my trivial perl questions), I thought it was once again time I bolstered my candidature for the useless brain-dead question of the year award (it makes a change from pressing for the equivalent award for answers/responses:-)), so here goes...

In my current incarnation, I've encountered numerous situations of late whereby a called method could do with establishing, not just the package/sub i.e. class method details, but the instance i.e. object, details of the caller - the currently vexatious situation revolves around the implementation of a dynamic bi-directional relationship between classes and indeed, objects.

To my (obviously inadequate) mind, the self-evident answers have, if not problems, then at the very least, limitations...

• Passing the caller class/instance as an argument looks not a little inelegant and also IMO, violates one, or more, OO rules & best practice(s).
• The use of inheritance is, apart from being frowned upon by the powers that be (since composition is the new inheritance), not applicable in all but one of the instances (pun intended)

• A manner similar to that of the automatically supplied called class/instance details or...
• An extension to caller() such that either the package or an instance (dependant on whether it was called by a package sub/class method or an object method) is returned as the 1st element of the list.

OTOH, maybe in addition to the above, the best solution would be to find/write a class to implement class/instance inter-relationships whereby each side of the relationship knows about both the...

• The class on the 'other side' of the relationship & the ...
• Arity of its own side of the relationship

.oO(...wonders whether if, in the light of 20/20 hindsight, this node might not be re-housed as a meditation - if the nodereaper can leave it alone that long:-)

Simple enough.

However, Foo::Module references Foo::Config, Bar::Module references Bar::Config, and each needs to talk to different databases, if nothing else, so the referenced modules need to remain distinct for each app's namespace, but I've had no luck in trying to track down a conventional way of accomplishing this.

What I've come up with so far is (in Base::Module):

BEGIN {
  my $ns = __PACKAGE__;
  $ns =~ s/::[^:]*$//;

  eval "require $ns\::Config";
  "$ns\::Config"->import(qw(x y z));
}

Is this the standard way of accomplishing what I'm trying to do or is there a better solution out there?

Hi all,

I'm about to start on a new web project using Perl. Currently I've Perl 5.10 installed on my computer, so my code will be tried and tested on this version. I've a feeling most hosting companies are still using Perl versions higher than 5 but below 5.10, so I think my code that's written for 5.10 should still run on these lower versions.

But will it run on Perl 6 when it becomes the default Perl installation? Will things like "use" or "autouse" still be used in Perl 6?

Thanks in advance!

if(@ARGV<4) {print "\nUse: SIZE START_X START_Y [0,1]" and exit;}
our ($N,$X0,$Y0,$VERBOSE) = @ARGV ;
our ($MOVES,@moves,@board)= (0,([-2,-1], [-1,-2], [-2 ,1],[-1 ,2], [2 
+,-1], [1, -2],[2, 1],[1, 2]),());
my $last_move = [$X0,$Y0] and inform_board();
$last_move  = inform_board($last_move) while($MOVES< $N*$N);
###############  SUBROUTINES ########################
sub init { return ($_[0]>$N/2)? init($N-1-$_[0],$_[1]) : ($_[1]>$N/2? 
+init($_[0],$N-1-$_[1]):($_[0]>=2 && $_[1]>=2? 8:(($_[0] >=2 && $_[1]=
+=1 || $_[1]>=2 && $_[0]==1)?6:($_[0]+$_[1]>1?4:($_[0]+$_[1]==1?3:2)))
+) )}
sub inRange{
my ($pos, $D) = @_;
my ($x,$y) = ($pos->[0]+$D->[0],  $pos->[1]+$D->[1]);
return  ($x>=0 && $x<$N && $y >= 0 && $y<$N && $board[$x][$y]>=0)?    
+ [$x,$y] : 0;
}
sub inform_board {
my ($moved_to) = @_;
if($moved_to)
{
$MOVES++;
print "\n$MOVES.",$moved_to->[0],",",$moved_to->[1] if $VERBOSE;
my ($MIN,$next_to_move) = (10,[]);
foreach(@moves)
 {
 if(my $finalPos = inRange($moved_to, $_))
  { 
  my $value_ref = \$board[$finalPos->[0]][$finalPos->[1]];
  ($MIN,$next_to_move, $board[$moved_to-> [0]][$moved_to->[1]]) =     
+($$value_ref,$finalPos,0) if(--$$value_ref>=0 && $$value_ref < $MIN);
+ } }
 return $next_to_move;
}
else
 {
 push @board, [(0)x $N] for(1..$N);
 for my $i (0..$N-1)  { $board[$i][$_] = init($i,$_) for (0..$N-1);}
  } }

Let's say I have these two bytes in a variable: 0xA8 0x38. ("Are these two bytes a string or a 16-bit integer?" you ask. I don't know; you tell me. This is part and parcel of my profound ignorance of this stuff. Pretend I just read the two bytes from a file into a variable.) These 16 bits are a packed representation of an MS-DOS FAT date.

• 5 bits 0–4 = Day (1–31)
• 4 bits 5–8 = Month (1 = Jan, 2 = Feb, etc.)
• 7 bits 9-15 = Year offset from 1980

I want to use Perl bitwise operations to convert this 16-bit integer (or string) to the date that it represents: 2008-05-08. Here's the conversion as I (naïvely) understand it:

A8       38         Hexadecimal (Little Endian)
10101000 00111000   Binary (Little Endian)

38       A8         Hexadecimal (Big Endian)
00111000 10101000   Binary (Big Endian)
11111100 00000000
54321098 76543210   Bit Ruler (Indexed From 0)

0011100 0101 01000  Binary (Regrouped)
1111110 0000 00000
5432109 8765 43210  Bit Ruler (Regrouped)

Year    Month Day   Packed Format

0011100 0101 01000  Binary
1C      05   08     Hexadecimal
28      5    8      Decimal

Year  = 1980 + 28 = 2008
Month = 5
Day   = 8 

MS-DOS FAT Date = 2008-05-08

Jim

O monks,

If I do  perl -e "require 'sys/ioctl.ph'" on most Linux systems, including Ubuntu Hardy, it works, but on Ubuntu Intrepid I get

Prototype mismatch: sub main::__LONG_MAX__ () vs none at /usr/lib/perl
+/5.10/_h2ph_pre.ph line 309.
Constant subroutine __LONG_MAX__ redefined at /usr/lib/perl/5.10/_h2ph
+_pre.ph line 309.

Is this just a bug in perl 5.10? Can anyone reproduce this in perl 5.10 on a non-ubuntu system? If it is a bug, what's the best way to report it?

TIA! -- Ben

my $self = bless {%${"$class\::Defaults"}}, $class;

I am making up a website that contains only static pages, using Template Toolkit. Original pages are in Dutch, but an English version is also required. So I will have template pages that look like this:

<p>Ik woon in Antwerpen.</p>

I need an easy way to rip these strings out of the xhtml-like templates (no worries if the odd span or b tag gets left in, but paragraphs should be intact), and writes them into a file(s) to be translated. The template also needs to be re-written as (something like) ...

<p>[% _tx(Ik woon in Antwerpen.) %]</p>

... so that TT can make the relevant substitution at the final pass

Now, I am far from the first to travel this road ... so I don't want to be making any new wheels. Does anyone know where I can borrow some old ones that work well enough?

Hi Monks :). OK, not sure if this is the appropriate section to ask, it seemed the only possibility to me given the options, and I do know as a newbie I've probably made some poor decisions in previous posts, but since my attempt is both sincere and dedicated, I accept my mistakes as par for the course, so please forgive if this is the case here.

I have a concern about the approach I'm taking in learning Perl, and was hoping to perhaps get some advice on whether or not the concern is warranted, one, and what to do if it is thought to be so, two. I am a mother, I work full time and have have quite a hectic life. I will begin school in summer 2009, however at the moment, I am self teaching, until I begin moving towards my bachelor's in comp science. As a result, I find myself, at times, going a week, sometimes even two!, without coding or studying at all. Granted there are plenty of times when I get stuck on a project, glued to the computer, my eyes glaze over, and someone has to literally come and pry me away, but as I said, I do find myself going a few days to a week or so without even looking at any of it.

Now please make no mistake, the dedication is there. I love coding and am fully prepared for all the good and bad it will come along with (my spouse is a developer, and he "goes into the basement" quite often). It's just that, in order to spend a substantial amount of time studying every day, I would have to seriously restructure my, and my children's, entire life, and I am willing to do this if needed, however I'm not sure where this stands in importance to my actual career, development as a hacker, ability, skill and any possible problems that may result from this approach down the line.

I know I can't be the first coder with a family, these worries, and this situation, so I thought I'd ask your opinion. Is a complete restructure of my approach a serious benefit in the pursuit of a career in this field, or is it kind of like riding a bike, once you advance to the next skill level, you're pretty much there? Of course we're all individuals and no one can answer this for me, or make the decision for me, I'm just curious because I do respect your opinions very much, as all of you seem to have achieved greatly in an extensive knowledge of the language.

Please note, I'm asking in concern about my eventual outcome, the type of coder my education is likely to produce, not so much about the time issue. I of course realize I'd have to make serious changes to get it done faster, but, again, this is not my focus at the moment. Also, please do not doubt my dedication because of this, I certainly have it abundantly. If I had two of me, the best one would be stuck to the computer non-stop, trust me! ;)

I don't think there's anything stunningly novel about it, but I found it to strike a good balance between capability and simplicity. It was pretty easy for other people to pick up and start writing tests within, and for diagnosing the inevitable test failures.

I wrote it before the recent TAPification of Perl test tools, so it could probably make use of some of the newer modules. (I wrote this tool partly in response to the difficult of extending the older Test::Harness to do what I needed.)

Enough introduction. Why should anyone care? I'll try to enumerate what seems to be unique or vaguely unusual about this tool:

• It is intended for processing the output of an application run on a given input file.
• It uses a simple but extensible syntax for describing the set of tests to run.
• The user can easily rerun specific failing tests with a very flexible way to pick the appropriate one. (This is slightly harder for this tool, since multiple tests can use the output of a single run.)
• The output checks are simple regular expression matches. You may have multiple checks per test. Tests can be either positive or negative (X must match and Y must not match.) Before release, I'd be tempted to make this a bit more flexible.
• Simple, TAP-compatible (I think) output
• Convenient for smoke tests. Our continuous integration tool ran a large set of these tests after every build, and only created a release package if all tests passed. If a test failed, it would say exactly what it wanted and didn't get, together with the actual output.

I'm meditating about a macrosystem which is working after compilation and before execution

The idea is to flag subs as :macro and and to replace every call with the string the sub returns.

use B::MyMacro;
sub foo :macro {
  return "bar()"
}

 
print;
foo();
print;

print;
do { bar() };
print;

Of course eval "string" needs also to be overridden, such that the macroreplacement will also be executed after eval's compilation.

This mechanism can already be realised using B::Deparse, RegEx and evals, but with rather fragile dependencies. I'm looking for a stable and simple opcode-based solution, replacing every macro-call with new opcodes.

IMHO, this approach would be much more stable and useful than codefilters, because "only perl can parse perl" ¹.

I'm wondering if nobody ever tried to implement this, googling and supersearching didn't bring any examples. Do you know any other similar approach?

UPDATES:
(¹) ... but "Even B:: can parse optree" : )