Automatically call ListingTableConfig.infer()

**Is your feature request related to a problem or challenge? Please describe what you are trying to do.**
A clear and concise description of what the problem is. Ex. I'm always frustrated when [...]
(This section helps Arrow developers understand the context and *why* for this feature, in addition to the *what*)

```rs
let ctx = SessionContext::new();
let s3 = object_store::aws::AmazonS3Builder::from_env()
.with_bucket_name("my-bucket")
.with_region("us-east-1")
.build()?;
ctx.runtime_env()
.register_object_store("s3", "my-bucket", Arc::new(s3));
let uris = vec![
"s3://a/b/c/1.parquet",
"s3://a/b/c/2.parquet",
];
let uris = uris.into_iter().map(|uri| ListingTableUrl::parse(uri).unwrap()).collect::>();
let table = Arc::new(ListingTable::try_new(ListingTableConfig::new_with_multi_paths(uris))?);
let df = ctx.read_table(table)?;
```

gives

```
error: Internal error: No schema provided.. This was likely caused by a bug in DataFusion's code and we would welcome that you file an bug report in our issue tracker
```

I had to manually call `infer()`:

```rs
let mut config = ListingTableConfig::new_with_multi_paths(uris);
config = config.infer(&ctx.state()).await?;
```

**Describe the solution you'd like**

DataFusion should call `infer()` if config is not set.
Parquet files have schema in them, and it is common to omit schema when reading or writing parquet files.

Start by reproducing the example through ListingTableConfig::new_with_multi_paths, ListingTable::try_new, and ctx.read_table, then inspect how ListingTableConfig::infer(&ctx.state()) is currently used. Done means a missing schema is inferred automatically for the shown Parquet paths and the example succeeds without the manual infer call.

Written by the indexing model from the issue text.