You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
This repository has been archived by the owner on Oct 23, 2019. It is now read-only.
Currently handling JS, JSON, CSS, and binary files all via the one epic pipeline definition.
Should break them up and try to find a simple and flexible way of resolving the types from the path or extension. Webpack's config mess should be avoided. Might be worth considering any benefits that can be drawn out of Ninja's concept of a low-level build description that is generated by code.
Need to consider how to handle compilation targets (eg: scss -> css). In a couple of projects I've had a pre-compiler that does some initial transformation and then indicates to the process that its output should be considered CSS. Does tend to get a bit messy when having to manually deactivate the postcss stuff.
Might be worth considering a fundamental assumption that all files may go through multiple transforms. Could get messy though, might end up with the same problems as browserify/webpack where they have elected for a low-level textual primitive that requires constant re-parsing for each transform.
The text was updated successfully, but these errors were encountered:
Currently handling JS, JSON, CSS, and binary files all via the one epic pipeline definition.
Should break them up and try to find a simple and flexible way of resolving the types from the path or extension. Webpack's config mess should be avoided. Might be worth considering any benefits that can be drawn out of Ninja's concept of a low-level build description that is generated by code.
Need to consider how to handle compilation targets (eg: scss -> css). In a couple of projects I've had a pre-compiler that does some initial transformation and then indicates to the process that its output should be considered CSS. Does tend to get a bit messy when having to manually deactivate the postcss stuff.
Might be worth considering a fundamental assumption that all files may go through multiple transforms. Could get messy though, might end up with the same problems as browserify/webpack where they have elected for a low-level textual primitive that requires constant re-parsing for each transform.
The text was updated successfully, but these errors were encountered: