Fork gulp tasks into gulp-sequelize and reintegrate.

[dkushner]

Since gulp has already been introduced as a dependency of the CLI binary, and most of the binary's functionality is based on a set of gulp tasks, I propose the creation of a separate gulp-sequelize project that implements these tasks in a way that conforms to gulp best practices.

After this has been completed, it should be relatively trivial to reintegrate with this standalone binary by simply making gulp-sequelize a package dependency and interacting with its tasks programmatically (as the binary already does).

The advantage of this is that it provides a set of functional tasks to developers who already have an existing gulp-based workflow while simplifying the development of the standalone binary.

[sdepold]

Sounds like a good idea. I pinged you on IRC. Let's discuss it there!

[sdepold]

Have you been able to actually fork the repository?

[maxmalov]

Hi, guys! Any updates on this?

[dkushner]

None from me, @maxmalov. I've been bogged down at work and also ran into some issues when I first gave it a shot. In my opinion, the best way to go about this would be to separate the actual migration logic (which is already sort of sequestered in the "helpers" directory) into some sort of sequelize-util repo or back into sequelize proper. This repo can then be a dependency of sequelize-cli, gulp-sequelize, grunt-sequelize, etc...

What do you think about this @sdepold? I'd be for rolling this functionality into sequelize proper. Performing migrations and managing models is inseparable from the ORM itself.

[sdepold]

Actually we will are currently planning to move the migration logic into a separate package.

[dkushner]

@sdepold, perfect! Is there an issue thread somewhere where we can collaborate on the API design for this move? I think Sequelize could benefit from some lightweight conventions, which I think you've already touched on by giving the CLI the ability to stub out models and migrations files a la Rails. I'd love to give my input as someone that's using Sequelize in a soon-to-be production environment.

[brad-decker]

@sdepold Is there a way to use the gulp tasks in our own gulp file currently? This is a great feature and will simplify build processes.

[sdepold]

Can you describe the setup. Which tool should use your own gulp file?

[brad-decker]

We want to run seeds on our test database for testing, and to run every time before Travis runs. We have a series of functions that run in order and the seed:create command is one.

For now just using a gulp plugin to run commands on shell.

[sdepold]

So the gulpfile itself is here https://github.com/sequelize/cli/blob/master/lib/gulpfile.js

[sdepold]

And the task is defined here https://github.com/sequelize/cli/blob/master/lib/tasks/seed.js#L21-L47

[sdepold]

You might be able to load the gulpfile and to use the respective task. but I'm not entirely sure how hard that is

[timini]

Has anyone managed to run any of the sequelize tasks from their local gulpfile?

This would be very useful, but at the minute I see is it blocked by the arg creating code in the bin/sequelize file.

[sdepold]

AFAIK nothing happened here.

[sushantdhiman]

480f3fc