WIP: Refactor server database #6263

By creating an object library for the server, we can more easily start unit-testing individual pieces of the code as we can simply link against that library instead of having to attempt to somehow link against the main executable (which is not really possible).

This commit should be removed once upstream Soci has integrated the necessary changes in their master branch. Then we should point the submodule there instead of to my forkThis commit should be removed once upstream Soci has integrated the necessary changes in their master branch. Then we should point the submodule there instead of to my fork..

The function for raising an error inside a custom trigger used to automatically roll back the currently active transaction. However, if the transaction is explicitly handled via a RAII wrapper that automatically rolls back on error, this would lead to a second (invalid) rollback command to be issued, causing an error. Instead, we now only abort the current sql statement for which the trigger failed, but we don't touch the active transaction.

That way, we have a (good) chance that seconds_since_epoch actually return the desired seconds since Unix epoch time and not just the seconds since the start of that particular clock.

Otherwise, the line number reported in case of a failing assertion won't be of any help as it will always point to the respective function definition instead of the function for which the assertion failed.

- Properly handle metagroups (e.g. "all", "in", "sub", ...) - Handle group modifiers (e.g. "!", "$", ...) - Set up default groups and ACLs when creating new server

The SuperUser is now no longer excluded from e.g. storing channel listeners in the DB. This should now make the handling of this special user more consistent with other parts of the code. Additionally, the SuperUser was handled specially only by chance as the user ID was checked to be != 0 to exclude the SuperUser. However, in general the Mumble::SUPERUSER_ID could have any arbitrary value.

When setting a volume adjustment for a non-existent channel listener via the Ice API an error will now be returned.

This happens e.g. if the registration happens via Ice for a user that is currently not connected. In this case, we simply default to using the root channel as the user's last channel.

Otherwise, each group could at most have a single member as adding a second user would violate the unique constraint on the (server_id, group_id) pair in the primary key.

When migrating a table, there might be tables that are new and that therefore do not exist yet. Furthermore, tables could have been renamed and therefore an old table with the same name doesn't exist either.

Scheme version 6 was introduced in 2014 and therefore we expect that any server that is still in use has been updated to a more recent version between then and now. If not, servers will have to be updated to any version >= 1.3.x which will perform the necessary migration and afterwards the DB scheme can be migrated by the current code. Maintaining backwards compatibility for more than that would only unnecessarily increase the maintenance burden (for testing the migrations).

This function has only very recently been introduced in SQLite 3.38.0 and is therefore not available in a lot of versions that are currently in use.

While we are not making any changes to the database, it is always good practice to also start a transaction for read-only DB accesses. Furthermore, BLOB objects can only be used while having an active transaction when using the PostgreSQL backend.

This ensures that the ensureTransaction function calls all play nicely together, avoiding instantiation of a duplicate transaction (which would lead to an error).

This makes the Binary data type slightly non-portable because it can only be sized for MySQL. However, there we actually need the sizedness in order to not end up with size-one binary objects, which would break our migration tests.

This allows to compare a DataType object for its type only - ignoring its size. Previously the comparison also compiled, but used the implicit construction of another DataType object and thus ended up comparing the size as well.

When using the BLOB approach as is done for MySQL and SQLite, Postgres will only insert the BLOB's OID into the column, instead of the BLOB's contents.

Each supported backend can be excluded from the build separately

The dumped JSON format may not be 100% DB backend independent but it should be somewhat close to it. Implements mumble-voip#4980 Implements mumble-voip#2960 (must be done manually though)

Mumble treats usernames case-insensitively. However, in the server, there is a cache that maps usernames to IDs of registered users. We do lazy loading so when a given name is not yet in the cache, we check the database for a match and if that returns something, we insert the result into the cache. However, the DB does case-insensitive lookups whereas the cache used to be case-sensitive. Therefore, different casings of the same username could end up in the cache (all pointing to the same user ID). If that user got unregistered, only one of the possible ways of writing their name gets removed from the cache, leaving the other entries (which are now stale) untouched. This can cause problems when another user now wants to register with a name that corresponds to one of those stale cache entries. The server will think that the given username is already taken, ensuring leading to a rejection of the registration request. A server restart purges the stale entries from the cache. This commit ensures that the cache is now also operating in a case-insensitive manner such that we shouldn't ever create any duplicate entries in there in the first place. Fixes mumble-voip#6383

The expected behavior is to return an empty string as the name. However, the current code will run into a database error that will shut the server down.

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

WIP: Refactor server database #6263

WIP: Refactor server database #6263

Commits on Aug 20, 2024

WIP: Refactor server database #6263

Are you sure you want to change the base?

WIP: Refactor server database #6263

Commits on Aug 20, 2024