If
you have a question on how to do something in Perl, or
you need a Perl solution to an actual real-life problem, or
you're unsure why something you've tried just isn't working...
then this section is the place to ask.
However, you might consider asking in the chatterbox first (if you're a
registered user). The response time tends to be quicker, and if it turns
out that the problem/solutions are too much for the cb to handle, the
kind monks will be sure to direct you here.
But, since Raisin is a functional framework, there's just a single application created, with the second "app" just being concatenated to the first, and with imported methods failing (eg, plugin) because they get imported into the first namespace that uses them.
I'm playing around with possible solutions, but nothing has stuck so far:
Reloading the entire Raisin::API module under a different namespace for each app (breaks helper methods in Raisin, so I'd have to rename all modules and code - way too messy)
Attempting to overload the "my" variables via an extra 'package Raisin::API' block (didn't think it would work, but worth a try - it didn't).
What I *think* I need is a way to load Raisin::API and it's dependencies in a scope limited way to lock it to each app's instance, but I can't think of how to do it. Would that be doable, or should I just drop the idea of trying to run multiple Raisin apps under Plack in this way? (yes, this is an odd setup, but it's transitional for a year or two while migrating away from a monolith)
Something like this:
my $app1 = some_scope_wrapper(sub {
use Raisin::API;
build_app1_here();
});
my $app2 = some_scope_wrapper(sub {
use Raisin::API;
build_app2_here();
});
# Raisin code won't work here...
$ cat test.pl:
#!/usr/bin/perl
use strict;
use warnings;
use POSIX qw( :unistd_h :sys_wait_h setsid WNOHANG );
POSIX::setsid() or die 'failed';
$ perlcritic --stern ./test.pl
Subroutine "setsid" not exported by "POSIX" at line...
What am I missing?
Perl::Critic has certainly found examples of dubious technique on my part, but it also seems to have a few blind spots.
What do the wise folks here think of Perl::Critic generally?
If someone has an example of a Pure PERL top-ish type command, pass it along.
Many of the LXD tools I've written would be far easier to use, if rewritten as top-ish commands, so this is the project I have in mind.
Thanks!
I'm using
use diagnostics
and very occasionally I get a
Use of uninitialized value in hash dereference at ....
message.
Basically, I have a background task that is accumulating information into a hash.
This hash is then transacted back to the parent.
I'm using Data::Dumper
to convert the hash in the background task into a string.
Once this string is received (by the parent), I then convert the string back into a hash with:
my %info = %{ eval $dumper };
And it is this line that generates the "uninitialized value" message.
(The background task sends data back to the parent every 10 seconds, and the "uninitialized warning" message is only genereted once in 2-3 weeks).
Before executing the line above, I do check that $dumper contians $VAR1=.
I'm struggling to find out what it is about the original hash that is being packed by Data::Dumper
that might cause this "uninitialized value" message.
One line of inquiry might be to see if I can get PERL to break-point when this "uninitialized value" message occurs ...
... at least then I could inspect the value of $dumper at the point of failure.
Any thoughts on how I might track down this intermittent error would be greatly appreciated.
Thank you.
Here's a quick solution I put together.
Feel free to improve or discuss.
Use of features from more recent Perls is fine (do indicate the version required).
Use modules if you want.
Golfing solutions are acceptable.
Update:
I received feedback from ++LanX#11140763 and
++NetWallah#11140770 regarding issues with my original code.
I made changes accordingly.
I also noted that the issue identified by NetWallah,
was also present in the next OUTER if $input->[$i] > $target; statement:
I've removed that line and added two more tests (which also show that zero is a valid target).
The new code follows.
The original code can be found at the end in the spoiler.
#!/usr/bin/env perl
use strict;
use warnings;
use constant {
INPUT => 0,
TARGET => 1,
EXPECTED => 2,
};
use Test::More;
my @tests = (
[[2,7,11,15], 9, [0,1]],
[[2,7,11,15], 9, [1,0]],
[[3,2,4], 6, [1,2]],
[[3,2,4], 6, [2,1]],
[[3,3], 6, [0,1]],
[[3,3], 6, [1,0]],
[[-5,-3,1,4,7], 1, [1,3]],
[[-5,-3,1,4,7], 1, [3,1]],
[[1,-1], 0, [0,1]],
[[1,-1], 0, [1,0]],
);
plan tests => 0+@tests;
for my $test (@tests) {
is_deeply
sort_arrayref(two_sum($test->[INPUT], $test->[TARGET])),
sort_arrayref($test->[EXPECTED]);
}
sub two_sum {
my ($input, $target) = @_;
my $got;
OUTER:
for my $i (0 .. $#$input - 1) {
for my $j ($i + 1 .. $#$input) {
if ($input->[$i] + $input->[$j] == $target) {
$got = [$i, $j];
last OUTER;
}
}
}
return $got;
}
sub sort_arrayref {
my ($aref) = @_;
return [ sort { $a <=> $b } @$aref ];
}
Hello folks - sorry, I thought I had got there after my last post, however I only got as far as accessing from a separate PL file. I'm now trying to ensure I can load the lexicon with the schema load and not everytime I call a method in my result / resultset classes (which seems like a really terrible idea).
So to try and give a complete picture, here's the script I eventually got to work:
#!/usr/bin/perl
use strict;
use warnings;
use FindBin qw( $Bin );
use lib "$Bin/../lib";
use Data::Dumper::Concise;
use TopTable::Maketext;
use Config::ZOMG;
use Path::Class::Dir;
# Load the Catalyst config
my $tt_config = Config::ZOMG->new( name => "TopTable" );
my $config_hash = $tt_config->load;
# Load the locales from the config
my (@locales, %inheritance, $config);
$config = $config_hash->{I18N}{locales};
foreach my $locale (keys %$config) {
push(@locales, $locale);
$inheritance{$locale} = $config->{$locale}{inherits} if defined $con
+fig->{$locale}{inherits};
}
# Get the directory where the messages are defined
my $dir = Path::Class::Dir->new( "$Bin/..", "root", "locale" );
# Load the lexicon
TopTable::Maketext->load_lexicon(
locales => \@locales,
directories => [$dir],
gettext_style => 1,
inheritance => \%inheritance,
);
my $lang = TopTable::Maketext->get_handle( "en_GB" );
printf "%s\n", $lang->maketext( "menu.title.league-tables", "Division
+Three" );
1;
use utf8;
package TopTable::Schema;
# Created by DBIx::Class::Schema::Loader
# DO NOT MODIFY THE FIRST PART OF THIS FILE
use Moose;
use MooseX::MarkAsMethods autoclean => 1;
extends 'DBIx::Class::Schema';
__PACKAGE__->load_namespaces;
# Created by DBIx::Class::Schema::Loader v0.07037 @ 2013-12-03 11:04:4
+4
# DO NOT MODIFY THIS OR ANYTHING ABOVE! md5sum:uMxbZipkwEqVJYByeZhY5Q
# You can replace this text with custom code or comments, and it will
+be preserved on regeneration
__PACKAGE__->meta->make_immutable(inline_constructor => 0);
1;
I'm very much a Moose novice I'm afraid, but believed if I added a 'lang' attribute with a builder method that sets all that up, I could then access that from my DB methods:
use utf8;
package TopTable::Schema;
# Created by DBIx::Class::Schema::Loader
# DO NOT MODIFY THE FIRST PART OF THIS FILE
use Moose;
use MooseX::MarkAsMethods autoclean => 1;
extends 'DBIx::Class::Schema';
__PACKAGE__->load_namespaces;
# Created by DBIx::Class::Schema::Loader v0.07037 @ 2013-12-03 11:04:4
+4
# DO NOT MODIFY THIS OR ANYTHING ABOVE! md5sum:uMxbZipkwEqVJYByeZhY5Q
use FindBin qw( $Bin );
use TopTable::Maketext;
has "lang" => (
is => "ro",
isa => "TopTable::Maketext",
builder => "_set_maketext",
required => 1,
);
sub _set_maketext {
my ( $self ) = @_;
my $class = $self->class;
my $app = $self->_app;
my (@locales, %inheritance);
my $config = $app->config->{I18N}{locales};
$app->log->debug( sprintf( "app: %s, class: %s", $app, $class ) );
printf( "app: %s, class: %s", $app, $class );
foreach my $locale (keys %$config) {
push(@locales, $locale);
$inheritance{$locale} = $config->{$locale}{inherits} if defined $c
+onfig->{$locale}{inherits};
}
my $dir = Path::Class::Dir->new( "$Bin/..", "root", "locale" );
TopTable::Maketext->load_lexicon(
locales => \@locales,
directories => [$dir],
gettext_style => 1,
inheritance => \%inheritance,
);
return TopTable::Maketext->get_handle( "en_GB" );
}
# You can replace this text with custom code or comments, and it will
+be preserved on regeneration
__PACKAGE__->meta->make_immutable(inline_constructor => 0);
1;
This however, doesn't work - the 'lang' method is accessible, but returns undef - I have added this as a test in one of my resultset methods:
$logger->( "debug", $self->result_source->schema->lang->maketext("menu.title.league-tables", "Division Three") );
But this gives an error:
[error] Caught exception in TopTable::Controller::Admin::Bans->process
+_form "Can't call method "maketext" on an undefined value at D:\Perso
+nal\Dev\Web\www.mkttl.co.uk\TopTable\lib/TopTable/Schema/ResultSet/Ba
+n.pm line 88."
Now Data::Dumper has the Freezer interface, which provides a way to modify the object before dumping - useful if the object caches a bunch of verbose but uninteresting stuff which will automatically be recreated when needed.
What Data::Dumper doesn't seem to provide is a way to say "call this method/function to get the _actual string_ this object should dump as". For this case it would be ideal to represent a GMP object $z as something like "z$z", taking advantage of its existing stringification overload - despite Data::Dumper's stated purpose, I hardly ever care about evalling the output to recreate the data structure.
Does anyone know of a recommended (or remotely usable) way to achieve such a thing? Ideal would be a per-class override, but even a global one would nearly always be as useful to me.
Update: I'm perfectly happy to use a module other than Data::Dumper, clearly that wasn't clear.
Update: I've managed to make some progress with Data::Printer after a few false starts. I'm going to spend some time seeing if I can learn to live with its output format (which has lots of options, but not necessarily the ones I'd want), but I like that I can throw almost everything in a config file so there's minimal boilerplate when I use it:
The hardest bit was avoiding its handling of duplicate references, which was making the last line of output show var[2] instead of z_1. That's what the unsee() call is fixing for me.
Snippets of code should be wrapped in
<code> tags not<pre> tags. In fact, <pre>
tags should generally be avoided. If they must
be used, extreme care should be
taken to ensure that their contents do not
have long lines (<70 chars), in order to prevent
horizontal scrolling (and possible janitor
intervention).
