fix: provide a proper error message when using getCloudflareContext in static routes

[dario-piotrowicz]

getCloudflareContext can't be used in static routes, currently a misleading error
message incorrectly tells the developer that they haven't called initOpenNextCloudflareForDev
in their config file, this change updates such error message to properly clarify what
the issue is (and how to solve it)

---

fixes #317

[changeset-bot]

🦋 Changeset detected

Latest commit: c93380e

The changes in this PR will be included in the next version bump.

This PR includes changesets to release 1 package

Name	Type
@opennextjs/cloudflare	Patch

Not sure what this means? Click here to learn what changesets are.

Click here if you're a maintainer who wants to add another changeset to this PR

[pkg-pr-new]

Open in Stackblitz

pnpm add https://pkg.pr.new/@opennextjs/cloudflare@330

commit: c93380e

[vicb]

Would there be a way to test this? (i.e. launch next build and check the errors`)

[dario-piotrowicz]

Would there be a way to test this? (i.e. launch next build and check the errors`)

I was thinking the same too, I don't think this could be tested using our e2es/playwright, so testing this would require a new type of test, a sort of semi-e2e one in which we run the build command and test the result, that's totally doable I think, but not trivial (but also probably not super complex) which would also increase maintenance costs. I am not sure if it's worth it (especially for just testing these error messages), but if you think it is I'm happy to work on it (potentially in a followup PR as there are other things that we could test in such a way, like the messages we present to the users during build)

just let me know can I can work on it 👍

[vicb]

Would there be a way to test this? (i.e. launch next build and check the errors`)

I was thinking the same too, I don't think this could be tested using our e2es/playwright, so testing this would require a new type of test, a sort of semi-e2e one in which we run the build command and test the result, that's totally doable I think, but not trivial (but also probably not super complex) which would also increase maintenance costs. I am not sure if it's worth it (especially for just testing these error messages), but if you think it is I'm happy to work on it (potentially in a followup PR as there are other things that we could test in such a way, like the messages we present to the users during build)

just let me know can I can work on it 👍

Thanks, a follow up PR sounds great, can you create an issue?

[dario-piotrowicz]

Thanks, a follow up PR sounds great, can you create an issue?

#331 🙂👍

(PS: again I'm not sure how we want to prioritize this, if you want me to work on it, I'm happy to work on it whenever)