Repository navigation
Conversation
)" This reverts commit a9b8bcc.
|
Recommendation: ready. WhyPull request #1190 put a draft RFC on WhatCommit The restored How
No commit between #1190 and current ImpactThe observable contract does not change. Decision 11 in This removes liability. Blocking findingsNone. Optional findingsNone in this diff. AlignmentDrafts #1204 and #1205 do not add the removed path. Their descriptions link the fork copy and say they stay drafts until the RFC is accepted. #1186 is a different served-optimize proposal ( #1201 and #1202 insert a registry row while the removed row is still hunk context ( No open pull request on this repository puts the RFC file back yet. That matches the split in this description: revert here, and reopen the text in a new pull request. #1175's description still names merged #1190 as the living RFC. That description is outside this diff. TradeoffThe proposal is absent from ValidationHead reviewed: Local, in a detached worktree at that head:
CI run 40914269203 is on this same head:
GQT regression proof does not apply. This pull request changes only RFC documents. It does not change query rows, counts, types, or errors. LimitationsI did not run the GQT corpus. I did not re-read the Lance compaction pages. This diff does not change storage code. I did check that
I read the open pull request list and the README diffs that can touch this registry row. I did not rebuild every open branch. |
…-optimize-13aba2 # Conflicts: # docs/rfcs/README.md
Reverts #1190 (a9b8bcc). The RFC was merged while it was still a draft under review; it should have stayed an open pull request. This restores the three RFC cross-references and the registry row to their pre-#1190 state and removes the draft document from
main. A fresh pull request re-opens the RFC text for continued review.No code change;
python3 scripts/check-docs.pypasses.