-
-
Notifications
You must be signed in to change notification settings - Fork 1.5k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
In memory build/serve #3494
Comments
This option has been added to support this workflow: https://rollupjs.org/guide/en/#watchskipwrite Otherwise, the existing hooks should be sufficient. You will not save a lot, though. |
@lukastaegert Thanks for that. Are there any other new developments in the perf area with rollup 2.0 ? Want to make sure I am doing the best practices for dev mode. My team's frustration with the build times is growing and I absolutely want to continue using Rollup. |
Unfortunately not, and dev performance is just one of a huge list of stuff I want to work, so no promises for any improvements soon. But if you really want a quantum leap in dev experience, maybe switching to a no-bundler-in-dev setup would work for you? There are quite a few tools that could make this work, but of course the devil is in the details.
So the question would be how much you depend on other Rollup plugin functionality. Things such as direct CSS imports might be fixable by adding a middleware to |
Thanks again for your reply. I ham aware of Snowpack and es-dev-server. I even published a middleware for Yeah these are great tools but unfortunately our build has some more parts like |
I am probably biased but I am really enjoying the None the less whenever you go to production you need an actual build and that is where rollup shines. Again if you only use what is possible in the browser then this part is a breeze 👍 I, however, can fully relate if that seems "impossible" for many cases as commonjs or other non standard, non browser features (like importing scss, css, ...) are heavily used. This sort of locks you into your build process where you can't easily switch to another build tool or have no build tool 😅 For commonjs you can sort of solve it by asking for es module versions or publish it automatically via for other special feature you will need to "pay the bill" by always requireing a build tool or by needing to reimplementn these features multiple times (for dev server, testing, building for production, ...) |
There would be ways to add a fast development experience to Rollup but as I said, there are quite a few other equally important topics I want to tackle first, and it would not be trivial. But yes, I think about performance a lot. |
I would love to help with that. In particular, looking for ways to include rollup in https://brunch.io. Brunch is one of the oldest web app builders around, it was there even before Grunt. I just can't stand complexity of every other bundler. Rollup is awesome for tree shaking. Brunch is very fast and simple. Snowpack, for example, requires some weird |
Rollup is being slow for me in development, it takes around ~20s for incremental builds in the 'watch' mode with cache.
Feature Use Case
If we could prevent outputting the incremental bundle to disk, instead serve it via another plugin (rollup-plugin-serve-inmemory ?). Just like webpack does. I think it will save considerable time for each iteration.
Feature Proposal
Do not output to disk, provide plugin hooks to enable development of a plugin which could serve the bundles from memory.
The text was updated successfully, but these errors were encountered: