hubcio commented on PR #2537:
URL: https://github.com/apache/iggy/pull/2537#issuecomment-3728550608
@seokjin0414
The tests you added are functional and work correctly, but they don't
address the core goal of #1657. Let me clarify what's needed.
The existing internal tests (`get_by_offset.rs`, `get_by_timestamp.rs`)
use a test matrix with **432 parameter combinations**:
- Message sizes: 50B, 1KB, 20KB
- Batch patterns: small/medium/large/very_large
- Segment sizes: 10B, 200B, 10MB
- Cache index configs: None, All, OpenSegment
- Messages required to save: 1, 24, 1000, 10000
Your PR uses fixed hardcoded values (`MESSAGES_COUNT = 2000`,
`BATCH_LENGTH = 100`) and doesn't vary any of these parameters. The whole point
of the issue is to get the samecomprehensive coverage through the public API.
What's needed:
1. Parameterized tests (or a loop-based approach to reduce server starts
as mentioned in the issue)
2. Variable message payload sizes matching the original tests
3. Adjustable server configuration for segment size, cache indexes, flush
thresholds
4. Same batch patterns as the internal tests
Keep in mind that if you do it one-to-one, we'll spawn around 1300 servers
and the test time would be in hours, so you have to optimize it (reuse
iggy-servers across multiple testcases).
Let me know if you have questions, happy to help on Discord.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
