Stabilize ghost-text size, token-budget completion length, tabby model lineup, full-accept reset#251
Merged
Merged
Conversation
Remove the explicit word-range cue from both the local-model and Apple Intelligence prompts so completion length is governed solely by the shared token budget (request.maxPredictionTokens). Bump suggestedPredictionTokenBudget 50% (11/18/30 -> 17/27/45) so the cap has room to land on a natural stopping point instead of hard-truncating mid-thought. The completionLengthInstruction parameter stays wired for a one-line revert.
Ghost font size is derived from the resolved caret height, but AX caret geometry is eventually consistent: the same field yields a tight line-height caret on one poll and the full field-height AXFrame fallback on the next. That made ghost text balloon whenever the coarse fallback won a poll. Track the smallest caret height seen during a focus session (keyed by FocusTracker's focusChangeSequence) and clamp larger readings down to it, so the suggestion stays the size of the real text line. The baseline resets on field switch or focus loss. The existing font-size floor bounds how small a spurious low reading can make the text; positioning is intentionally unchanged. Collapse presentOverlay's loose geometry parameters into the FocusedInputContext they already came from, which also threads focusChangeSequence through to the overlay.
The token budget for .twelveToTwenty was bumped to 45, but this test still expected the old 30 and was failing on main. Update it to match the intended budget.
…gth cue - Model display names back to tabby-fast-1 / tabby-balanced-1; drop the Qwen3.5 entry (unrecognized series), prefer gemma-4-E2B then Qwen3-0.6B - Add a Reset-to-default (backtick) button to the full-accept keybind in settings and onboarding, so it can't get stuck disabled - Re-add the completion-length cue to the Foundation Model prompt while the llama path stays token-budget only - Stop tracking .writing/, remove the old Tabby->Cotabby rename docs
This was referenced May 25, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Validation
Risk / rollout notes