From: "Andrew Moore" <[EMAIL PROTECTED]>
Date: 31 October 2008 12:46:06 AM
To: "Adam Thick" <[EMAIL PROTECTED]>
Subject: Re: [Koha-devel] [RFC: DBIx::Class 2/3] DBix::Class -
moving C4::ClassSource over to using DBIx::Class
Hi Adam -
Thanks for taking the time to look over the patch and make your
comments. I'm glad that you agree that something like this would help
Koha. I'm hopeful that I can get it refined enough to get it accepted.
I think you and I agree on a lot of the points you brought up. Let me
hit a few of them one at a time:
On Wed, Oct 29, 2008 at 7:47 PM, Adam Thick
<[EMAIL PROTECTED]> wrote:
I also think it would be a really good fit for Koha to
standardise the calls to the database and ensure the SQL is
sanitised.
I'd love to one day see Koha as a set of objects, like Koha::Member
or
Koha::Biblio, where Koha has attributes that can be handed round to
each
object, and each object with it's own set of attributes and methods
to get
things done, ah but one day!
Yes, I'd like to make many of the C4::* packages into classes of
objects, and probably inherit properties from a base class, too. I
think that's a step that will be easier to take after this one,
though. Joe Atzburger and I have been playing with a few different
possibilities, so stay tuned for proposals along those lines. If you
have an example implementation of this, I encourage you to send it
along to the koha-devel list to gauge the community response.
I'd recommend creating the DBIx::Class schema with detailed column
information and inflator/deflators for things like dates and such
as a start
too. This is time consuming to go back and add later, and having
this kind
of information helps create SQL from your schema and provide
upgrade scripts
:)
That's a good point. In the improvements that I've made since Friday,
I'm using the Schema::Loader to make files that represent each of the
tables, instead of loading them to memory each time. In them, I'm
inflating dates, but nothing else. What else should I be inflating?
my $sources = GetClassSources();
@@ -91,13 +94,11 @@ foreach my $cn_source (sort keys %$sources) {
sub GetClassSources {
my %class_sources = ();
- my $dbh = C4::Context->dbh;
- my $sth = $dbh->prepare_cached("SELECT * FROM
`class_sources`");
- $sth->execute();
- while (my $source = $sth->fetchrow_hashref) {
- $class_sources{ $source->{'cn_source'} } = $source;
+
+ my @class_sources = $schema->resultset('ClassSources')-
>search();
+ foreach my $row ( @class_sources ) {
+ $class_sources{ $row->get_column('cn_source') } = {
$row->get_columns };
}
- $sth->finish();
return \%class_sources;
I found rather than using arrays to store the search in it was a
bit faster
( cause your not loading up an array of data ) and syntactically
prettier to
get a result set like this
my $class_sources = $schema->resultset('ClassSources')->search();
while ( my $row = $class_sources->next() ) {
$class_sources->{$row->cn_source} = $row->get_columns
#not sure if this can be called on a cursor or not, might want to
look at
something like
http://search.cpan.org/~ash/DBIx-Class-0.08010/lib/DBIx/Class/ResultClass/HashRefInflator.pm
I'm not a DBIC expert, but I tried sometime similar to that at first
and then switched to my method for some reason. I think it was so that
I could get ahold of the ->columns call. I was probably overlooking a
call like the one you show. I'm sure that we'll get better at using
DBIC as we go along. I'll depend on folks like you to help us use it
better.
I am really, really happy that database abstraction is being
considered for
Koha. It will mean the code base will be easier to maintain and can
potentially make database upgrading easier too. I'd like to know if
you are
considering a MVC approach to Koha ( in the long run :) I've been
running
with Catalyst for a while now too and find that it's great for
separation
between Controller, Model and View and definitely makes maintenance
and
coding easier. Separating out the code like this also means we can
re-purpose elements like the Model to perform other tasks like
providing a
webservice :) or cron scripts etc...
I'd like to switch to something like Catalyst at some point, but
that's a pretty long-term goal. If you have any ways to help move in
that direction, I encourage you to share them with us. The roadmap
that I'm working on kind of goes along the lines of:
* abstract out the SQL using DBIC or maybe SQL::Abstract or
DBIx::Abstract
* improve the database upgrade tools by relying on something like
SQL::Translator and DBIx::Class::Schema::Versioned
* start making packages into classes. Maybe use something like Moose?
* Move to something like Catalyst or whatever is good by this point
in time.
I'm glad that you and I seem to be in agreement on the kinds of things
that would make Koha easier to use, maintain, and extend. Please don't
hesitate to contribute your thoughts (or patches!) to the koha-devel
list.
Thanks!
-Andy
