These are a bit more questionable but still keep things cleaner a bit, at least in some cases?
Most of these were the result of `cargo clippy --fix -- -W clippy::use_self`
To keep with DRY principle, I think it makes it a bit less redundant to reuse the Self::<associated_type_name> structure in the well known trait implementations - keeps it consistent with the trait decl too.
- Disables HTTPs support by default, the TLS stack is opt-in
- Renames "http_client" feature to "http-client"
- Uses native TLS by default in pyoxigraph and cli
- Uses Rustls for Linux Python wheels and Docker images
Uses by default targets of 1GB of RAM and 2 threads.
Data parsing is in most of the case slower than ingestion so no more than 2 threads are used anyway.
With read-only it's not possible to modify the data. Updates to the data
are possible via a primary instance of oxigraph, but will not be
reflected. The data is frozen at the time the read-only server is
started.
The database can be opened once in primary mode, but may be opened
multiple times in secondary mode. It's not possible to write data to the
database in secondary mode.
Secondary mode does not support checkpoints so there might be data
inconsistencies when data is changed while said data is queried. It
might happen that the result data mixes both elements before and after
the change that can not happen in primary mode.
Yuri Astrakhan