Module cloud_sources 

Every object store datui can read, as a list of sources rather than one of each kind.

A source is a store plus the login that reaches it: the default S3 settings, a Google login, or an entry in [[cloud.connections]]. Each has an ID, and the ID is what keeps two stores apart when their bucket names collide: two MinIO servers can both have a bucket called data, so a URL from an S3-compatible source names it, s3://<id>@bucket/key. Everywhere else the location is unambiguous and URLs stay the standard ones, so a URL copied out of datui still works in any other tool.

Discovery here reads environment variables and asks whether files exist. It never touches the network: listing is cloud_browse’s job, and it runs on a worker.

Structs

Resolved

A cloud URL resolved to the plain URL the libraries understand and the settings that reach it.

S3Settings

How to reach one S3 or S3-compatible store.

Source

One store and the login that reaches it.

Enums

Signing

Whether requests to a place carry a signature.

Tier

Where a source came from. Lower wins when the same ID turns up twice, and sources are listed in this order.

Constants

DEFAULT_AZURE_ENV

An Azure account named in the environment: a connection string, or an account with a key or SAS token.

DEFAULT_AZURE_LOGIN

A signed-in az.

DEFAULT_GCS

The Google login found in the environment or the application-default file.

DEFAULT_S3

The source built from [cloud] s3_*, --s3-* and the AWS_* environment: what a plain s3://bucket/key has always meant.

Functions

access_key

The bucket or container url is in, as one string: s3://bucket, s3://<id>@bucket, gs://bucket, abfss://container@account. What is learned about signing is kept per place.

discover

Every source this machine and the config describe, in display order.

endpoint_host

The host and port of an endpoint URL, for display.

expand_azure_short_url

An az://, adl:// or azure:// URL, container/path, as its canonical abfss:// form. These name no account, so it comes from where the URL was typed (inside an account on the home screen), the environment, or the one account in the config. Every other path comes back as it is.

is_within

Whether url is root or somewhere inside it. Azure URLs are compared in their canonical form, and a trailing slash does not matter.

known_access

What this session learned about signing requests to the place url is in: true for unsigned.

on_home

The sources the home screen shows: every [[cloud.connections]] entry, and of the logins found on the machine, the kinds [cloud] discover allows.

profile_source_id

The ID of the source for an AWS profile: aws- and the profile’s name, lowercased, with anything that cannot go in an ID turned into -.

remember_access

Remember for the session whether place (an access_key) is read unsigned.

remember_bucket

Remember that source reaches bucket, for URLs that do not name their source.

remember_listed

Buckets an earlier run listed, remembered as if listed now: a bucket under Recent opens with the login that found it before its source is listed again. Only S3: Google’s first level is projects, and an Azure URL names its account.

resolve

Resolve url against the sources in config and on this machine, with an Amazon S3 bucket’s own region. May run a credential command and ask S3 where the bucket is, so call it on a worker.

resolve_for_open

As resolve, for opening an object. A place whose signing is still Signing::Try is settled first with one unsigned request, since the libraries that open it make many requests and cannot retry them without a signature.

resolve_with

As resolve, with the environment supplied.