fix: skip possible api route when ENOENT

[leoj3n]

Description

Before PR:

After PR:

Additional context

Related file/line:

vite/packages/vite/src/node/server/transformRequest.ts

Lines 111 to 113 in 2351d79

	if (e.code !== 'ENOENT') {
	throw e
	}

Not sure if this is the right location and code to fix this but it does resolve the issue in current state.

Not sure if test is needed or where would be added.

What is the purpose of this pull request?

• Bug fix

Before submitting the PR, please make sure you do the following

• Read the Contributing Guidelines.
• Read the Pull Request Guidelines and follow the Commit Convention.
• Check that there isn't already a PR that solves the problem the same way to avoid creating a duplicate.
• Provide a description in this PR that addresses what the PR is solving, or reference the issue that it solves (e.g. fixes #123).
• Ideally, include relevant tests that fail without this PR but pass with it.

[patak-dev]

Thanks for the PR. For the moment, I don't think we can proceed with it. We will move away from suffixes once some proposals advance. This is not only an issue for raw but for other suffixes too. For the moment, I think the best is to add a middleware in your project to handle these routes before they get to Vite's pipeline. See https://vitejs.dev/guide/api-plugin.html#configureserver

[leoj3n]

@patak-dev Thanks for the explanation. Can you point to the specific proposals? Interested to take a look.

[patak-dev]

See this issue #7017, there are links from there to the diff proposals.

[leoj3n]

See this issue #7017, there are links from there to the diff proposals.

@patak-dev What will take the place of suffixes? AFAICT the linked issue just swaps assert for as proposal but doesn't seem like that affects suffix API in any way?

[patak-dev]

We are going to get something like this in the future:

import foo from './file' as 'raw'

[leoj3n]

We are going to get something like this in the future:

import foo from './file' as 'raw'

@patak-dev Q is regarding suffix on URLs (http://localhost:3000/foo?raw), not the reflective import in files (see OP).

We will move away from suffixes

What this entails does not seem explained in #7017. Did you mean "move away from asserts"?

What will take the place of suffixes?

Maybe I'm not understanding but I don't get how using either reflective import or assert would affect the URL suffix.

[patak-dev]

We aren't going to use the ?raw suffix anymore if we manage to use as 'raw', not only for glob imports but in general. Support for ?raw could be removed in a future major.

[leoj3n]

@patak-dev Ah, okay. If you care to, can you explain why ?raw was needed in the first place and why it is obviated by the move from assert to as? I assume I could figure it out if I were to study the code or wait for ?raw to be removed and see what changes then but am curious now.