Closed izaera closed 6 years ago
If we dropped gulp as the CLI and task runner, does that mean we would also drop the use of streams and gulp plugins?
Probably, but not necessarily. It would be a matter of internal implementation that shouldn't be exposed to the users. I mean that users shouldn't care (or take advantage) of our internal implementations.
For example, as it is done now, there are a lot of internal tasks and subtasks which are difficult to understand (some are obvious, some don't) and they are -in fact- an API, because people can be defining custom builds and attaching to them, for example. I'm not against a gulp API, but definitely not one that exposes internal tasks that we may want to remove, rename or change.
However, take this with a pinch of salt, as I'm thinking aloud, not asserting what I think must be done.
Nevertheless, do you think gulp streams are so necessary? I used other similar tools for liferay-npm-bundler (like when copying files and things like that) and they look fast too 🤔 .
Closing here. We're for now still under gulp. We can rethink this inside liferay-themes-sdk if this ever becomes relevant again.
Currently this tool is structured as gulp tasks which has some drawbacks like:
On the other hand, it is already written using gulp and it's working, so we should see a clear benefit to stop using it. Also, if we stop using gulp, we need to:
Which looks like a considerable effort.
So, I would like to use this issue to discuss pros and cons of:
Any ideas?
/cc @Robert-Frampton @jbalsas