4paradigm / 4paradigm/OpenMLDB
deprecating legacy schema codec
- Dominant language
- C++
- Stars
- 1.7k
- Forks
- 331
- Avg merge
- 12d 12h
- Merged PRs (30d)
- 1
Description
The schema codec in https://github.com/4paradigm/OpenMLDB/blob/ff7e8acf21ead1f9734eef59ac3521adb05dff2e/hybridse/include/codec/fe_schema_codec.h#L34-L110
is not real serialization solution for schemas. It
1. Not extendable. complex data type like #3680 is not possible without breaking change, it can not represent map type in single byte.
2. Hardware dependent. Endianness and primitive type size vary on different platform
3. Lack of type safety check, and is error prone.
## Solution ?
Contributor guide
Research direction
The issue points to hybridse/include/codec/fe_schema_codec.h lines 34-110, which contains the legacy schema codec. First, examine the current codec implementation to understand its limitations. Then, look for any existing serialization solutions in the codebase that could replace it. The goal is to design a new, extensible, platform-independent, and type-safe schema serialization method, but the issue does not specify a concrete solution or next steps.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- cpp
- Domain
- backend, databases
- Issue type
- Refactor
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 30/100