fix: calculate header percentage based on available input space

[roomote]

Follow-up to #11034

Fixes the percentage calculation in the collapsed task header to show the correct value.

Problem

The percentage in the header was showing 45% when it should show ~19%. For example, with:

• 50.4k tokens used
• 128k reserved for response
• 400k context window

The old formula calculated: (tokensUsed + reservedForOutput) / contextWindow * 100 = (50.4k + 128k) / 400k * 100 = 44.6% (incorrect)

Solution

The new formula calculates: tokensUsed / (contextWindow - reservedForOutput) * 100 = 50.4k / (400k - 128k) * 100 = 18.5% (correct)

This represents the percentage of available input space that is currently used. The reserved output tokens should not be counted towards the used percentage since they are reserved for the model response.

Changes

• Updated the percentage calculation in TaskHeader.tsx to use the correct formula
• Added edge case handling for when available input space is 0 or negative
• Added a test to document the expected behavior

Testing

• All existing tests pass
• Added new test case for percentage calculation logic

---

View task on Roo Code Cloud

---

Important

Fixes percentage calculation in TaskHeader.tsx to correctly compute available input space usage, with updated tests in TaskHeader.spec.tsx.

• Behavior:
  • Fixes percentage calculation in TaskHeader.tsx to use tokensUsed / (contextWindow - reservedForOutput) * 100.
  • Handles edge cases where available input space is 0 or negative.
• Testing:
  • Adds test cases in TaskHeader.spec.tsx for new percentage calculation logic.
  • Ensures tests cover edge cases and correct percentage display.
• Misc:
  • Updates comments and documentation in TaskHeader.tsx to reflect new calculation logic.

^{This description was created by for 51e062c. You can customize this summary. It will automatically update as commits are pushed.}

[roomote]

Rooviewer   See task on Roo Cloud

All issues have been resolved. The test now properly mocks useSelectedModel and getModelMaxOutputTokens with known values and verifies the rendered percentage matches the expected calculation.

• Test does not verify percentage calculation - add assertions for the actual rendered percentage value

Previous reviews

• d56b3da: Review #1
• d56b3da: Review #2

_{Mention @roomote in a comment to request specific changes to this pull request or fix all unresolved issues.}

[roomote]

Rooviewer   See task on Roo Cloud

Review complete. The core formula fix is correct, but the test needs improvement.

• Test does not verify percentage calculation - add assertions for the actual rendered percentage value

_{Mention @roomote in a comment to request specific changes to this pull request or fix all unresolved issues.}

[roomote]

Fixaroo   See task on Roo Cloud

Fixed the test to properly verify percentage calculation by mocking useSelectedModel and getModelMaxOutputTokens with known values and asserting the rendered percentage. All local checks passed.

_{View commit | Revert commit}

[hannesrudolph]

Tested it. Works