Optimize binary file size

[hnwyllmm]

Description

The current stripped release binary for seekdb is approximately 380 MB, which is excessively large. Since the binary file is loaded into memory at runtime, this also impacts memory usage.

Proposed Optimization Areas

The following table outlines potential areas for code removal or optimization to reduce binary size:

Optimization Point	Action / Details	Notes / Considerations
General
Remove Oracle mode code	Delete
Remove shared storage code	Delete
Remove redundant virtual tables	Delete tables like GV$, CDB_*
Charset	Keep only utf8mb4 and related collations	No requirement for AP/AI scenarios; needs evaluation for TP scenarios.
Backup/Restore/Archive/CDC	Move to close_module	CDC functionality must be retained.
RPC Code	Convert to local calls if only synchronous invocation is used. Delete serialization/deserialization code for objects only used in synchronous calls.	Note: Some DDL RPCs may be invoked from thread pools.
gcc_except_table	Remove most exception tables and stack info (~28 MB). SEEKDB does not use exceptions for error handling (except for objit).	Some exception handling may need to be retained.
GIS Code Optimization	~33.9 MB. Contains many template cross-redundant codes (e.g., class Comparison<A,B> and class Comparison<B, A>).
TABLE API	Delete
Vector Engine Optimization	Remove Oracle-related code. Optimize comparison functions, etc.: remove infrequently used types and convert them to slow-path calls. Clean up redundant templates like func(a,b) and func(b,a). Convert non-performance-critical type support (e.g., JSON) to slow path (requires specific analysis).
Direct Load	Non-core functionality; evaluate for deletion.
PX/PDML/DTL	Delete code related to remote nodes.
Unit Manager	Delete
Multi-tenancy