felipecrv commented on code in PR #40325:
URL: https://github.com/apache/arrow/pull/40325#discussion_r1513026171
##########
cpp/src/arrow/filesystem/azurefs.cc:
##########
@@ -65,6 +65,133 @@ AzureOptions::AzureOptions() = default;
AzureOptions::~AzureOptions() = default;
+Result<AzureOptions> AzureOptions::FromUri(const arrow::internal::Uri& uri,
+ std::string* out_path) {
+ AzureOptions options;
+ const auto host = uri.host();
+ std::string container;
+ std::string path;
+ if (arrow::internal::EndsWith(host, options.blob_storage_authority)) {
+ options.account_name =
+ host.substr(0, host.size() - options.blob_storage_authority.size());
+ auto components = internal::SplitAbstractPath(uri.path());
+ if (!components.empty()) {
+ container = components[0];
+ path = internal::JoinAbstractPath(components.begin() + 1,
components.end());
+ }
+ } else if (arrow::internal::EndsWith(host, options.dfs_storage_authority)) {
+ options.account_name =
+ host.substr(0, host.size() - options.dfs_storage_authority.size());
+ container = uri.username();
+ path = uri.path();
+ } else {
+ options.account_name = uri.username();
+ std::string host_port = host;
+ const auto port_text = uri.port_text();
+ if (!port_text.empty()) {
+ host_port += ":" + port_text;
+ }
+ options.blob_storage_authority = host_port;
+ options.dfs_storage_authority = host_port;
+ if (uri.scheme() == "abfs") {
+ options.blob_storage_scheme = "http";
+ options.dfs_storage_scheme = "http";
Review Comment:
So let's figure out a way to infer we're talking to Azurite, but not assume
that people using `abfs://` are OK with plain-text communications and would
remember to intentionally opt-in to security by using `abfss://`. Security
should be opt-out, not opt-in. If someone ever complains about `https://` by
default we should spec a URI query parameter to disable TLS instead of having
it subtly disabled by the scheme chosen.
Even the Microsoft documentation says that they will use TLS even when
`abfs` is used in the Haddop driver.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
