chore(jsonrpc): updated rpc_errors_schema

[miraclx]

Update rpc_errors_schema.json after the introduction of TooManyFunctions in #4954

[chefsale]

Don't have contaxt, but looks fine to me.

[matklad]

Couple of questions there:

• should we have a test to make sure the thing stays in sync?
• that error is part of a nightly protocol feature, if we end up removing or changing it in any way, will we be able to revert this change easily? As in, does this schema affect anything outside of nearcore which might be hard to change?

[frol]

that error is part of a nightly protocol feature, if we end up removing or changing it in any way, will we be able to revert this change easily? As in, does this schema affect anything outside of nearcore which might be hard to change?

@matklad Why that error variant is not behind a feature-flag then?

This file is used on near-api-js side to generate some "strictly typed errors"

[miraclx]

• should we have a test to make sure the thing stays in sync?

I'll add a check in CI that fails if there's a conflict in the schema

• As in, does this schema affect anything outside of nearcore which might be hard to change?

As far as I know, it's used in https://github.com/near/near-api-js/blob/ee975a1edd84dff955302453fbcca1413d362bb7/fetch_error_schema.js#L8, but so far, it's tagged to a specific revision - 4c11499, we can revise it before updating the file on that end.

[matklad]

Why that error variant is not behind a feature-flag then?

The code that produces error is gated, so gating the error variant itself isn't strictly necessary. Doing that is possible, it's just requires a bunch of extra cfgs everywhere. I and @Longarithm looked at that and figured that there's little benefit in that. No strong opinion on my side here though, could add a feature-gate there.

[miraclx]

Superseded by #5052