Batch sn payments

[darcys22]

No description provided.

[majestrate]

this will be a chonky PR when it's ready

[darcys22]

Todo:

• Tests
• switch sqlite_orm to another lib
• use trigger to clear out zero amount balances
• constrain sqlite table to only allow positive balances
• check functionality of subaddresses
• mining after SN network running causes miner to crash

Trigger

CREATE TRIGGER batch_payments_delete_empty
AFTER UPDATE ON batch_payments FOR EACH ROW WHEN NEW.amount = 0 THEN DELETE FROM batch_payments WHERE address = NEW.address;

Constraint

CHECK(amount >= 0)

[jagerman]

switch sqlite_orm to another lib

-> https://github.com/SRombauts/SQLiteCpp now used in storage server, lets you write your own queries but keeps a C++ interface that's much nicer to work with than the straight C interface.

[darcys22]

Misc other issues:

• Clean up CMake folder
• Rebase
• Simulate batching and decide on final length of batching period

[darcys22]

This was something that came up when trying to run cmake with -DCMAKE_BUILD_TYPE=Debug this made it work but unsure exactly what is going on behind the scenes and whether it should stay here or not

[jagerman]

-ldl is already linked for most things via the ${CMAKE_DL_LIBS}, included as part of the extra target dependencies. I'm guessing there's something not depending on extra that needs this? Whatever it is, it's better to do it as:

target_link_libraries(target INTERFACE ${CMAKE_DL_LIBS})

(where target is whatever the target is with the linking issue).

[darcys22]

This is now ready for review.

• The macos tests are failing and are awaiting an upstream change which should resolve the issue. Disable std::filesystem on macOS when compiling C++17 but targetting <10.15 SRombauts/SQLiteCpp#335
• There oxen_service_nodes_test_rollback test fails but is inconsistent, went over the codebase with Valgrind to see if I could find a memory leak/uninitialised variable which resulted in Initialise uptime proof info #1482 however I'm not sure if this was the root cause. Will have another go searching for this
• Static builds are failing but I've exhausted my knowledge trying to get them to pass and need some help. CMake is hard
• Discussion on optimal batching interval size to be scheduled

[jagerman]

First pass, left many comments/suggestions, some tiny, some bigger.

More to come.

[jagerman]

-ldl is already linked for most things via the ${CMAKE_DL_LIBS}, included as part of the extra target dependencies. I'm guessing there's something not depending on extra that needs this? Whatever it is, it's better to do it as:

target_link_libraries(target INTERFACE ${CMAKE_DL_LIBS})

(where target is whatever the target is with the linking issue).