Error when deploying Gen 1 Firebase function with no region set in config

Environment

Nuxt 3.7.1 with Nitro 2.6.2

Reproduction

Hard to provide a reproduction since deployment is necessary to reproduce.

I have tested with older versions to find the root cause, and it seems that providing no region in firebase config will cause the error. Once I provided a region the error no longer occured.

Describe the bug

When deploying, all you will see in console is Error: An unexpected error has occurred.

After digging into the firebase-debug.log, you will see this error:

[debug] [2023-09-06T00:35:12.299Z] TypeError: Cannot read properties of null (reading 'match')
    at resolveString (/usr/lib/node_modules/firebase-tools/lib/deploy/functions/params.js:31:27)
    at /usr/lib/node_modules/firebase-tools/lib/deploy/functions/params.js:47:36
    at Array.map (<anonymous>)
    at Object.resolveList (/usr/lib/node_modules/firebase-tools/lib/deploy/functions/params.js:47:21)
    at toBackend (/usr/lib/node_modules/firebase-tools/lib/deploy/functions/build.js:180:30)
    at Object.resolveBackend (/usr/lib/node_modules/firebase-tools/lib/deploy/functions/build.js:73:23)
    at process.processTicksAndRejections (node:internal/process/task_queues:95:5)
    at async prepare (/usr/lib/node_modules/firebase-tools/lib/deploy/functions/prepare.js:68:62)
    at async chain (/usr/lib/node_modules/firebase-tools/lib/deploy/index.js:38:9)
    at async deploy (/usr/lib/node_modules/firebase-tools/lib/deploy/index.js:95:5)

Obviously this is not very helpful. After some trial and error, I discovered that the functions.yaml output was slightly different between v2.6.2 & v.2.5.2:

v2.6.2:

{
    "endpoints": {
        "server": {
            "platform": "gcfv1",
            "availableMemoryMb": null,
            "timeoutSeconds": null,
            "minInstances": null,
            "maxInstances": null,
            "ingressSettings": null,
            "serviceAccountEmail": null,
            "vpc": null,
            "region": [
                null
            ],
            "httpsTrigger": {},
            "entryPoint": "server"
        }
    },
    "specVersion": "v1alpha1",
    "requiredAPIs": []
}

v2.5.2:

{
    "endpoints": {
        "server": {
            "platform": "gcfv1",
            "availableMemoryMb": null,
            "timeoutSeconds": null,
            "minInstances": null,
            "maxInstances": null,
            "ingressSettings": null,
            "serviceAccountEmail": null,
            "vpc": null,
            "httpsTrigger": {},
            "entryPoint": "server"
        }
    },
    "specVersion": "v1alpha1",
    "requiredAPIs": []
}

As you can see, the older version did not have "region" defined at all while the new version has it set to [null] by default. After discovering this I set the region in the nitro config and sure enough it deployed without the same error.

Additional context

No response

Logs

No response

Start by tracing how the deployment output functions.yaml is generated, comparing Nitro 2.6.2 with 2.5.2 and the region handling described in the report. Verify the Gen 1 Firebase deployment path when no region is configured; done means it no longer emits a [null] region or produces the reported Firebase error.