Open sheerun opened 10 years ago
I'm not sure that modifying Bowerfile.lock is a good idea. We are to have Bowerfile
and Bowerfile.lock
firmly synchronized. If some modifications we made to one file we should keep up-to-date another one. This topic need to be discussed more, I think.
I'll look more deeply into sprockets
I don't talk about modyfying Bowerfile.lock, but paths in sprockets :)
But it's possible there are many matching Bowerfile.lock for one Bowerfile.
For example if in Bowerfile we have "jquery", "> 1.0.0", the 1.1 and 1.2 versions in Bowerfile.lock are ok.
Sample use case:
Given User starts rails application with `rails server`
When User issues `bowered update` in separate terminal
Then Sprockets paths in running server are updated and assets re-compiled
And User sees updated assets after refresh in browser
Sample use case
Got it, thanks.
This should be partly extracted from bower-rails.
bowered install
programmatically beforeassets:precompile
I leave research for this task for @SergeyKishenin
Ideally there should be no need for restarting server in development for reloading assets.
Some cases to be handled:
Integration would be independent of rails, but could be hooked-up to
rails server
command if rails environment was detected (runbowered install --watch
as a thread fromrails server
command).We need to do research about changing sprockets paths in runtime.