SYNOPSIS
use aliased 'DBIx::Class::DeploymentHandler' => 'DH';
my $s = My::Schema->connect(...);
my $dh = DH->new({
schema => $s,
databases => 'SQLite',
sql_translator_args => { add_drop_table => 0 },
});
$dh->prepare_install;
$dh->install;
or for upgrades:
use aliased 'DBIx::Class::DeploymentHandler' => 'DH';
my $s = My::Schema->connect(...);
my $dh = DH->new({
schema => $s,
databases => 'SQLite',
sql_translator_args => { add_drop_table => 0 },
});
$dh->prepare_deploy;
$dh->prepare_upgrade({
from_version => 1,
to_version => 2,
});
$dh->upgrade;
DESCRIPTION
DBIx::Class::DeploymentHandler is, as its name suggests, a tool for
deploying and upgrading databases with DBIx::Class. It is designed to
be much more flexible than DBIx::Class::Schema::Versioned, hence the
use of Moose and lots of roles.
DBIx::Class::DeploymentHandler itself is just a recommended set of
roles that we think will not only work well for everyone, but will also
yield the best overall mileage. Each role it uses has its own nuances
and documentation, so I won't describe all of them here, but here are a
few of the major benefits over how DBIx::Class::Schema::Versioned
worked (and DBIx::Class::DeploymentHandler::Deprecated tries to
maintain compatibility with):
* Downgrades in addition to upgrades.
* Multiple sql files files per upgrade/downgrade/install.
* Perl scripts allowed for upgrade/downgrade/install.
* Just one set of files needed for upgrade, unlike before where one
might need to generate factorial(scalar @versions), which is just
silly.
* And much, much more!
That's really just a taste of some of the differences. Check out each
role for all the details.
WHERE IS ALL THE DOC?!
DBIx::Class::DeploymentHandler extends
DBIx::Class::DeploymentHandler::Dad, so that's probably the first place
to look when you are trying to figure out how everything works.
Next would be to look at all the pieces that fill in the blanks that
DBIx::Class::DeploymentHandler::Dad expects to be filled. They would be
DBIx::Class::DeploymentHandler::DeployMethod::SQL::Translator,
DBIx::Class::DeploymentHandler::VersionHandler::Monotonic,
DBIx::Class::DeploymentHandler::VersionStorage::Standard, and
DBIx::Class::DeploymentHandler::WithReasonableDefaults.
$dh->prepare_version_storage_install
Creates the needed .sql file to install the version storage and not the
rest of the tables
$dh->prepare_install
First prepare all the tables to be installed and the prepare just the
version storage
$dh->install_version_storage
Install the version storage and not the rest of the tables
WHY IS THIS SO WEIRD
DBIx::Class::DeploymentHandler has a strange structure. The gist is
that it delegates to three small objects that are proxied to via
interface roles that then create the illusion of one large, monolithic
object. Here is a diagram that might help:
The nice thing about this is that we have well defined interfaces for
the objects that comprise the DeploymentHandler, the smaller objects
can be tested in isolation, and the smaller objects can even be swapped
in easily. But the real win is that you can subclass the
DeploymentHandler without knowing about the underlying delegation; you
just treat it like normal Perl and write methods that do what you want.
THIS SUCKS
You started your project and weren't using
DBIx::Class::DeploymentHandler? Lucky for you I had you in mind when I
wrote this doc.
First, define the version in your main schema file (maybe using
$VERSION).
Then you'll want to just install the version_storage:
my $s = My::Schema->connect(...);
my $dh = DBIx::Class::DeploymentHandler->new({ schema => $s });
$dh->prepare_version_storage_install;
$dh->install_version_storage;
Then set your database version:
$dh->add_database_version({ version => $s->schema_version });
Now you should be able to use DBIx::Class::DeploymentHandler like
normal!
LOGGING
This is a complex tool, and because of that sometimes you'll want to
see what exactly is happening. The best way to do that is to use the
built in logging functionality. It the standard six log levels; fatal,
error, warn, info, debug, and trace. Most of those are pretty self
explanatory. Generally a safe level to see what all is going on is
debug, which will give you everything except for the exact SQL being
run.
To enable the various logging levels all you need to do is set an
environment variables: DBICDH_FATAL, DBICDH_ERROR, DBICDH_WARN,
DBICDH_INFO, DBICDH_DEBUG, and DBICDH_TRACE. Each level can be set on
its own, but the default is the first three on and the last three off,
and the levels cascade, so if you turn on trace the rest will turn on
automatically.
DONATIONS
If you'd like to thank me for the work I've done on this module, don't
give me a donation. I spend a lot of free time creating free software,
but I do it because I love it.
Instead, consider donating to someone who might actually need it.
Obviously you should do research when donating to a charity, so don't
just take my word on this. I like Matthew 25: Ministries:
http://www.m25m.org/, but there are a host of other charities that can
do much more good than I will with your money. (Third party charity
info here:
http://www.charitynavigator.org/index.cfm?bay=search.summary&orgid=6901
POD ERRORS
Hey! The above document had some coding errors, which are explained
below:
Around line 91:
Unknown directive: =method
Around line 98:
Unknown directive: =method
Around line 105:
Unknown directive: =method