Conversation
…etup (v3.5.86) Ghost text: after each foreground turn the auxiliary model drafts the line the user is most likely to type next; it renders dim at the prompt and Tab (or the right arrow) accepts it in full. Drafting runs on a background daemon thread, is silent on any failure, is one-shot per prompt, and is superseded rather than overwritten when a newer turn starts. New config key input_suggest (default true); CHEETAH_SUGGEST=0 disables it for one run. The prediction is applied synchronously at pre_run and on every text change because prompt_toolkit only consults the auto-suggester on text insert — an empty prompt is never asked, and a fast Tab can beat the async pass. Tab-title setup: the first-run helper only ever considered a local editor install, so a Remote-SSH / WSL / devcontainer session wrote (and fabricated) ~/.config/Code/User/settings.json on the server — a file the editor never reads — then marked itself done and never retried. It now locates the editor server install and writes the remote Machine settings the window actually reads, requires an existing User directory before touching the local path, prints paste-ready instructions when neither is reachable, and scopes its marker per target so legacy markers self-heal with one retry. Tests: 48 new cases (tests/test_input_suggest.py incl. end-to-end over a real prompt_toolkit session; tests/test_vscode_setup.py — the module had none). Full suite: 2629 passed, 5 skipped.
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
…etup (v3.5.86)
Ghost text: after each foreground turn the auxiliary model drafts the line the user is most likely to type next; it renders dim at the prompt and Tab (or the right arrow) accepts it in full. Drafting runs on a background daemon thread, is silent on any failure, is one-shot per prompt, and is superseded rather than overwritten when a newer turn starts. New config key input_suggest (default true); CHEETAH_SUGGEST=0 disables it for one run.
The prediction is applied synchronously at pre_run and on every text change because prompt_toolkit only consults the auto-suggester on text insert — an empty prompt is never asked, and a fast Tab can beat the async pass.
Tab-title setup: the first-run helper only ever considered a local editor install, so a Remote-SSH / WSL / devcontainer session wrote (and fabricated) ~/.config/Code/User/settings.json on the server — a file the editor never reads — then marked itself done and never retried. It now locates the editor server install and writes the remote Machine settings the window actually reads, requires an existing User directory before touching the local path, prints paste-ready instructions when neither is reachable, and scopes its marker per target so legacy markers self-heal with one retry.
Tests: 48 new cases (tests/test_input_suggest.py incl. end-to-end over a real prompt_toolkit session; tests/test_vscode_setup.py — the module had none). Full suite: 2629 passed, 5 skipped.