This will reduce the dependency load on libspark and the c++ algo sdk.
This has been on the backburner for a while; we're doing it now so we
can deliver a clean sdk to Laguna.
Migrating the following code:
* general server-side md utils (MarketDataGateway, etc.)
* emumba dependent code, including md impl for:
- deribit
* ccapi dependent code, including md impl for
- binance
- binance_us
- binance_usds_futures
- ftx
- coinbase
- okx
- deribit
* md recording
Note: I first began this migration to eliminate the c++ algo sdk's
wependency on emumba. There seemed to be a conflict related to ssl
versions (the build in the sdk docker would fail with unresolved
external symbols for OPENSSL macros)
Not sure how this slipped through the cracks, but I'm guessing the build
broke when we upgraded the cpp compilers (to clang 17 and gcc 13).
* Include <cstdint> in a few places
* Suppress -Wstringop-overflow in sparkd-gcc-release build
* Change const char * param to explicit std::string for proper type
deduction in call to 'std::make_tuple'. For some reason, the gcc build
was failing with an array-bounds werror (changing the arg fixed it).
The AddColumnsStream, SortStream, and requisite yacc/lex grammar were
originally added to a new branch, off of development, which is on net8.
I forgot to change it to net7 when I cherry-picked the commits for the
regular development branch.
It's wasn't quite as simple as flipping net8 -> net7 in csproj files,
because I was eager to use new language features only available in net8!
Namely, primary constructors and collection expressions.
Bugs (fixed):
* MarketDataClient.price_ticker, and price_stats streams don't fire update
with initial state (they only tick as subsequent changes are observed)
Good riddance. Improper 'PlatformMap' config was often the reason market
data mysteriously wouldn't show up in the client. Especially with new
exchanges. We should keep all md source/connection configuration in one
place (one proto message), to avoid any confusion.