Open shadowcat-mst opened 10 years ago
bjoernfan
commented
10 years ago I haven't really inspected the code, but what's the reason to generate .sql files? Why not use DBIx::Class::Schema->deploy? (Just sharing what I've used, not saying that it's better than any other solution).
I haven't used ::Candy either, but I'm using deploy here if anybody wants an example.
shadowcat-mst
commented
10 years ago On Mon, Apr 14, 2014 at 08:33:00AM -0700, Björn Pettersson wrote:
I haven't really inspected the code, but what's the reason to generate .sql files? Why not use DBIx::Class::Schema->deploy? (Just sharing what I've used, not saying that it's better than any other solution).
Because deploy is a one-off deploy - great for development, sure.
However the proposed tools are designed to handle versioning and upgrades too.
The reason to generate .sql files is so (a) people can actually read the SQL files (b) you can double check the results are sane (c) if you have any oddities to deal with you can edit them (and the code will use the files).
Plus, the diff files need something to diff against to generate them, and of course the old version of the schema isn't there anymore unless you ship a complete git repository with the app :)
Basically, ->deploy was the original one that I wrote - and is still great for "dump current schema version into the db" - but the more recent stuff is rather more sophisticated.
(also, I assumed that Schema::Loader meant that you had people who really really wanted to see SQL, because that's about the only way it makes sense for a new project that owns the database like PearlBee does :)
Matt S Trout - Shadowcat Systems - Perl consulting with a commit bit and a clue
http://shadowcat.co.uk/blog/matt-s-trout/ http://twitter.com/shadowcat_mst/
Email me now on mst (at) shadowcat.co.uk and let's chat about how our CPAN commercial support, training and consultancy packages could help your team.
bjoernfan
commented
10 years ago Wow, that was a patient answer! The more you know. Thank you. :)
Given you're using DBIx::Class anyway, why not use DBIx::Class::Candy to declare the result classes, then DBIx::Class::Migration / DBIx::Class::DeploymentHandler to generate the .sql files?
SQLite would make a great choice for small blogs and for experimenting with the software, and PostgreSQL is going to be a vastly preferred 'real' deployment option to MySQL for lots of people.
Plus ::Candy is so much prettier :D