Co-authored-by: ellipsis-dev[bot] <65095814+ellipsis-dev[bot]@users.noreply.github.com>
Co-authored-by: Matt Rubens <mrubens@users.noreply.github.com>
Co-authored-by: Daniel Riccio <ricciodaniel98@gmail.com>
Co-authored-by: Daniel <57051444+daniel-lxs@users.noreply.github.com>
* fix: resolve Claude Code token counting inefficiency and enable caching (#5104)
- Remove 1.5x fudge factor from Claude Code token counting
- Enable prompt caching support for all Claude Code models
- Add comprehensive tests for token counting and caching
- Update existing tests to reflect accurate token counting
This fixes the extreme token inefficiency where simple messages would
jump from ~40k to over 60k tokens, causing API hangs when approaching
the artificial 120k limit. Claude Code now properly utilizes its full
200k context window with accurate token counting.
* fix: address PR review comments
- Extract IMAGE_TOKEN_ESTIMATE as a named constant for clarity
- Update token counting tests to use exact counts instead of ranges for deterministic testing
- Fix test expectations to match actual tokenizer output
* Remove token counting changes, keep only cache support
- Removed custom countTokens override from claude-code.ts
- Deleted claude-code-token-counting.spec.ts test file
- Kept cache token collection and reporting functionality
- Kept supportsPromptCache: true for all Claude Code models
- Kept claude-code-caching.spec.ts tests
This focuses the PR on enabling cache support without modifying token counting behavior.
* fix: update webview test to expect supportsPromptCache=true for Claude Code models
---------
Co-authored-by: Daniel Riccio <ricciodaniel98@gmail.com>
Co-authored-by: Bruno Bergher <me@brunobergher.comexport>
Co-authored-by: Matt Rubens <mrubens@users.noreply.github.com>
Co-authored-by: Daniel Riccio <ricciodaniel98@gmail.com>
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
Co-authored-by: Daniel <57051444+daniel-lxs@users.noreply.github.com>
Co-authored-by: Daniel Riccio <ricciodaniel98@gmail.com>
Co-authored-by: Matt Rubens <mrubens@users.noreply.github.com>
* Fixes#4882: Remove experimental setting for command execution in attempt_completion
- Remove DISABLE_COMPLETION_COMMAND from experiments system
- Permanently disable command execution in attempt_completion tool
- Update tool prompts to remove command parameter and examples
- Remove experimental UI toggle and localization entries (18+ languages)
- Update tests to reflect permanent behavior
- Remove experiment-specific test file
Command execution is now permanently disabled in attempt_completion.
Users must use execute_command tool separately before attempt_completion.
* refactor: simplify getAttemptCompletionDescription by removing unnecessary variables
* test: fix tests by regenerating snaps
---------
Co-authored-by: Daniel Riccio <ricciodaniel98@gmail.com>
The test was expecting isFetching to be false in the default state, but the implementation correctly initializes it as true to show a loading state on initial load. Updated the test expectation to match the actual behavior.
Add support for the save/discard flow for support prompt setting page
Normally when you edit things on the settings pages, the save button lights up,
allowing you to discard your changes. Currently the prompts page doesn't support this flow-
the prompts are immediately saved when they change.
With this change, we use the normal cachedState system in the SettingView, allowing users to dicard changes to their prompts like any other setting.
This removed the need for the resetSupportPrompt event since we send the entire state of the support prompts (same as before).
Test plan:
* Manually verified prompts can be saved/discarded for different types of support prompts.
