It builds the anticipation

Following the general idea.

It's close to what I had hacked together before, but implementing this minimal version of it I think will help to build up the rest of the functionality more cleanly.

Still curious if I'm missing something re: `QueryCache.build` vs `QueryClient.setQueryData` @jumr.dev

Cool, that looks simple enough, but why do you use `QueryCache.build` instead of `QueryClient.setQueryData`? It looks like an undocumented API (at least I don't see it in the docs).

Thanks! I'd love to see an RFC as well. I'll see if I can distill my use case better...

I'm also entirely open to the suggestion that this is not a use case well suited to the library at this time.

The biggest challenge so far has been shipping to the app stores, since some users insist on it. iOS still has inconsistent support for PWAs and wrapping the site in a thin native client has presented some issues, particularly around SSO flows.

But I think the PWA route is serving us well. We only need to implement a new feature once, and it's immediately available to all users. It's also forcing us to consider responsive design for all new work, which, frustratingly, is seemingly still a lot to ask for these days.

Bare RN. We were on .59 and couldn't upgrade without first upgrading React, but we couldn't upgrade React without first upgrading React Router, etc.

We managed those migrations over the last 12 months thankfully. I might take another look at modern RN (I like that the bridge is gone)...

Is there a canonical source to understand this preference?

Are Icon and Tooltip 3rd party components? No, they're actually located in `a/b/c/components/`!

I'm sure this must have saved the original maintainers whole seconds when writing import statements by hand, but it is now costing me my sanity 10 years later.

The work consists of deleting the rule and then working my way through hundreds of now unresolved import errors and updating `import w from "z"` to `import w from "x/y/z"`.

I'm currently several weeks (part time) into removing our `resolve.modules` rule where we told webpack to look in `a/b/` for our modules... And also `a/b/c/`, and `a/b/c/d`, and `a/b/e`, and `a/b/f`, and...

The Get Started link on the landing page assumes that Tanstack Start has its own documentation, but it seems to still be a descendant of the Tanstack Router. Here's a working link tanstack.com/router/lates...

You didn't compute anything

Was this on the @sentry.io launch week schedule?