Replies: 5 comments 9 replies
-
@matthewp the main problem I found when using edge adapters is the fact that the edge runtime is not the same as the node runtime. However, there’s two contexts: building astro and running astro. It’s perfectly fine to use node apis and node c++ addons in the prior, not the latter. One of the great benefits of why I chose astro was that I could deploy it anywhere. Perhaps, the authoring integrations docs could describe the edge runtime limitations. Even better, it would be great if astro could detect if an integration (or even user code!) is using node apis and c++ addons in the astro middleware/astro component head to warn them of their impending incompatibility if they set their adapter to one of the edge deployment adapters. Detecting the usage of node apis and c++ addons that don’t exist would tell users immediately if their code is incompatible and save integration maintainers’ time. For example, for my internationalization integration, I had to disable three other edge-incompatible integrations to debug the underlying issue in my integration. |
Beta Was this translation helpful? Give feedback.
-
I've took a brief look at Astro codebase, and I have a concern about the possibility of core migrating into NodeJS ecosystem so far, it would become incompatible with Deno runtime. I'm not implying that it's gonna happen, but in case it does, I don't want to abandon an adapter after a breaking update comes out. You may actually remember my willing to work on Netlify Edge adapter, it is still there, but to me it looks like only the core team knows where Astro is going to go in the future. I'm sure that somewhere on these pages is a lot of useful insights, sorry, didn't find what I was hoping to see. Can we be sure that the core won't require Node-specific packages or system modules in the future? I think a have a few good ideas about what to add to the Netlify adapter, but first some help with getting into maintainers "position" would be appreciated Btw, using Astro for around 6 month, absolutely love it's SSG capabilities and developer experience. Clients absolutely love the speeed too |
Beta Was this translation helpful? Give feedback.
-
Could it make sense to also open dedicated Discord channels for dev discussions, which only relate the specific adapter? |
Beta Was this translation helpful? Give feedback.
-
i dont understand why vercel or astro dropped the vercel/edge part of the adapter and only support the middleware edge, |
Beta Was this translation helpful? Give feedback.
-
Same with Netlify lol. Welp, now we don't use Astro for SSR 🤷♂️ |
Beta Was this translation helpful? Give feedback.
-
Body
Summary
Find new maintainers for a few of the adapters and have core focus more on Astro core. Core will keep the Node adapter, because it is the most widely used, and the Vercel adapter, since they are the project's official host sponsor.
Background & Motivation
As Astro has grown the popular of SSR has increased, but we have been unable to keep up on issues with adapters. Additionally there are specific features that each adapter/host has that the core team doesn't necessarily have the knowledge to take advantage of.
The community has done a great job taking on the maintenance of many of these things and we want to officially hand over ownership. Additionally some 3rd party adapters such as SST adapter are among the best we have.
Goals
Non-goals
Beta Was this translation helpful? Give feedback.
All reactions