Commit Graph

316 Commits

Author SHA1 Message Date
Taylor Mullen c4c11f1d65 Prevent flickering on confirmation decline.
- When larger confirmations were shown and then declined you'd typicaly get large chunks of content flickering upon typing or sending a subsequent request. This was primarily due to us leaving the latest confirmation as "updateable" / pending. This changeset addresses that by flushing any pending confirmation to the static container.

Part of https://b.corp.google.com/issues/414196943
2025-05-13 16:36:28 -07:00
Jacob Richman e665d4f198
First step refactoring InputPrompt (#335) 2025-05-13 16:23:14 -07:00
Miguel Solorio c4fb1ad04b
Use `Enter` on slash commands to execute (#334) 2025-05-13 16:08:12 -07:00
Olcan 4a0f5476c0
fall back to ~/.env if .env is not found in current directory or any ancestors (#338) 2025-05-13 15:36:34 -07:00
Brandon Keiji 3be8b6dc34
fix(sandbox): default to current user profile for debian/ubuntu env (#337) 2025-05-13 21:13:54 +00:00
Olcan 0ae59056d1
mention `git diff --staged` for partial commits (#336) 2025-05-13 14:06:14 -07:00
Brandon Keiji 0e61a15438
fix: remove gemini-code bin configuration (#330) 2025-05-13 13:09:58 -07:00
Olcan 17cce0adf4
use git diff HEAD instead of git diff && git diff --staged (#333) 2025-05-13 12:56:32 -07:00
Jacob Richman e2c3611c63
Multiline editor (#302)
Co-authored-by: Taylor Mullen <ntaylormullen@google.com>
2025-05-13 11:24:04 -07:00
Brandon Keiji 8da7a71d9a
refactor: shorten 'gemini' binary name (#329) 2025-05-13 10:49:45 -07:00
Miguel Solorio 61ccd4f33a
Allow users to cancel out of theme selector (#310) 2025-05-13 07:41:32 -07:00
Taylor Mullen 7d8392bab4 Encourage utilization of `git diff --staged` for commit message writing.
- This enables GC to base commit message changes on actual file content.
2025-05-12 23:51:37 -07:00
Taylor Mullen 3217576743 feat: Enhance `replace` tool reliability with multi-stage edit correction
This commit significantly improves the `replace` tool's robustness by introducing a multi-stage correction mechanism. This directly addresses challenges with LLM-generated tool inputs, particularly the over-escaping of strings sometimes observed with Gemini models, and other minor discrepancies that previously led to failed edits.

The correction process is as follows:
1.  **Targeted Unescaping:** The system first applies a specialized unescaping function to the `old_string` and `new_string` to counteract common LLM-induced escaping patterns.
2.  **LLM-Powered Discrepancy Resolution:** If a unique match for the `old_string` is still not found, the system leverages a Gemini model (`gemini-2.5-flash-preview-04-17`) to:
    *   Identify the most probable intended `old_string` in the file by intelligently correcting minor formatting or escaping differences.
    *   Adjust the `new_string` to correspond with any corrections made to the `old_string`, maintaining the original edit's intent.

This enhancement makes the `replace` tool more resilient and effective, leading to a higher success rate for automated code modifications. The `expected_replacements` parameter has been removed as the tool now focuses on finding a single, unique, and correctable match. The tool's description and error reporting have been updated to reflect these new capabilities.

Fixes https://b.corp.google.com/issues/416933027
2025-05-12 23:33:12 -07:00
Olcan 5ec254253f
ensure no 'undefined' in system prompt (#322) 2025-05-12 16:41:11 -07:00
Olcan 884d6ebfd8
system prompt for working with git (#321) 2025-05-12 16:27:07 -07:00
Taylor Mullen df74594b8f When an error occurs stop processing. 2025-05-12 00:06:20 -07:00
Taylor Mullen 8537aabba4 feat: Add User-Agent to API requests
This change introduces a User-Agent header to all API requests made by the Gemini CLI.

The User-Agent string includes the CLI version, Node.js version, operating system, and architecture. This will help in tracking usage and identifying potential issues.

Fixes https://b.corp.google.com/issues/416353675

Signed-off-by: Gemini
2025-05-11 14:33:58 -07:00
Taylor Mullen 2970f0a06c feat: Integrate centralized error reporting for API interactions
Implements robust error handling for Gemini API calls, integrating with the centralized error reporting system.

- API errors are now caught and reported to dedicated log files, providing detailed diagnostics without cluttering the user interface.
- A concise error message is surfaced to the user in the UI, indicating an API issue.
- Ensures any pending UI updates are processed before an API error is displayed.

This change improves our ability to diagnose API-related problems by capturing rich error context centrally, while maintaining a clean user experience.

Signed-off-by: Gemini <YourFriendlyNeighborhoodAI@example.com>
2025-05-11 13:55:55 -07:00
Taylor Mullen 4d5f0dc080 Workaround model bug where it returns invalid history items.
- Currently there's a bug in the API (or SDK?) where the SDK endpoint will commonly fail with:

```
Error: Failed to generate JSON content: got status: 400 Bad Request. {"error":{"code":400,"message":"* GenerateContentRequest.contents[5].parts: contents.parts must not be empty.\n","status":"INVALID_ARGUMENT"}}
```

- At times the model will respond with an empty parts list where if we send that back up to the API endpoint it explodes with the above. Using a curated history seems like a total hack around this prolbem, and even in the SDK (i'm following up on this), BUT helps mitigate this issue.
2025-05-11 13:01:54 -07:00
Taylor Mullen cf91f72c5c Remove terminal tool and dependencies.
- We now solely use the shell tool. This deletes all content around the legacy terminal tool so we can focus on improving the new Shell tool.
- Remove instances from sandboxing, tests, utilities etc.
2025-05-11 12:35:55 -07:00
Taylor Mullen dcb67c32a5 Log server information on error.
- The goal of this is to enable us to better diagnose server issues when they occur.
- Added tests because why not.
2025-05-10 14:18:23 -07:00
Taylor Mullen d159a1507e Don't prematurely end convo w/ Gemini.
- There seems to be a root model bug where the model will preemptively bail on conversations without trying harder. Typically the stops are VERY obvious and bug-looking where you need to prmopt the model to "continue".
- This PR attempts to fix the above by running a 2.0-flash request (don't need somethign more powerful) at the end of every full interaction to see who should speak (user or model).
- Add tests for nextSpeakerChecker

Fixes https://b.corp.google.com/issues/416826051
2025-05-10 14:05:58 -07:00
Taylor Mullen c0eab31c02 Show model decline/cancellation states.
- Upon decline / cancellation we weren't showing the model the cancellation status or states. Therefore it didn't know why things would or wouldn't happen

Fixes https://b.corp.google.com/issues/416797704
2025-05-10 00:26:18 -07:00
Taylor Mullen 6b518dc9e4 Enable tools to cancel active execution.
- Plumbed abort signals through to tools
- Updated the shell tool to properly cancel active requests by killing the entire child process tree of the underlying shell process and then report that the shell itself was canceled.

Fixes https://b.corp.google.com/issues/416829935
2025-05-10 00:21:09 -07:00
Taylor Mullen 090198a7d6 Make cancel not explode.
- We were console.erroring, throwing and early aborting. Instead we now treat cancels like a normal user message and show an indicator in the UI

Fixes https://b.corp.google.com/issues/416515841
2025-05-09 22:49:32 -07:00
Taylor Mullen 28f9a2adfa fix: Resolve infinite loop
- This change addresses and resolves an infinite loop. The patch ensures the loop condition is correctly handled, preventing its recurrence.
- Added tests for markdownUtilities.test.ts

Fixes: https://b.corp.google.com/issues/416795337

Signed-off-by: Gemini <My circuits hummed, and the loop was no more.>
2025-05-09 17:37:36 -07:00
Allen Hutchison 4a6d0717a1
fix for b/414940078 (#306) 2025-05-09 15:38:19 -07:00
Taylor Mullen e9274b2ab2 feat: Update default Gemini model to gemini-2.5-pro-preview-05-06
Fixes https://b.corp.google.com/issues/416778280

Signed-off-by: Your Witty AI Assistant
2025-05-09 15:27:51 -07:00
Olcan 92c1279de6
sandbox/seatbelt-aware system prompt, support for custom seatbelt profiles under project settings (#304) 2025-05-09 11:33:05 -07:00
Taylor Mullen b8fa38a6e8 feat: Improve theme not found handling
Modify  to return a boolean instead of throwing an error when a theme is not found. Update CLI startup and  hook to handle the boolean return value for more graceful error handling.
2025-05-09 10:28:20 -07:00
Olcan c58f879026
fix MCP under seatbelt, improve error handling (#301) 2025-05-09 09:02:14 -07:00
Olcan b35a3856a2
fix debugging with seatbelt, including in strict profile (#300) 2025-05-09 08:44:40 -07:00
Taylor Mullen baa26e9e2e Ensure dogfood packaging more accurately mirrors local build.
- Removed `build:package` in favor of `npm run build`.
- The regular build does extra work to copy over relevant information into the `dist` dir. Alternatively without this we get a `dist` dir in the `cli` folder that has no seatbelt packaging.

Fixes https://b.corp.google.com/issues/416634356
2025-05-08 23:55:44 -07:00
Taylor Mullen 41b82ce796 Add bundling support.
- This can now be invoked with `npm run bundle`, it creates a `bundle/` folder that has:
  - gemini.js
  - sandbox-macos-minimal.sb
  - sandbox-macos-strict.sb
  - shell.json
  - shell.md

- This doesn't include any sort of automation for auto bundling pieces. It's just the root capability which we can weave into other locations.

Fixes https://b.corp.google.com/issues/411432723
2025-05-08 23:36:42 -07:00
Amir Hardon 1c486a4050 Fix: Prevent CLI from crashing when a configured theme is not found
Previously, if a theme specified in the user's settings was not found, the CLI would crash during startup. This was particularly affecting users upgrading from older versions as the "ANSI colors only" theme was renamed to "ANSI".

This commit adds error handling to catch the theme not found error during initial loading and when setting themes later. Instead of crashing, the application now logs a warning, displays an error message in the UI, and opens the theme selection dialog to allow the user to choose a valid theme.
2025-05-08 22:33:46 -07:00
Brandon Keiji 4741c9a6eb
fix(sandbox): set --inspect-brk in production sandbox when env DEBUG is truthy (#295) 2025-05-08 21:12:19 -07:00
Miguel Solorio a685597b70
UI Polish for theme selector (#294) 2025-05-08 16:00:55 -07:00
Olcan 6b0ac084b8
allow SEATBELT_PROFILE=none to disable seatbelt on macos (#296) 2025-05-08 15:52:04 -07:00
Olcan b1c449d11c
refined sandbox/seatbelt log message, pass NODE_OPTIONS along to sandboxed node (#292) 2025-05-08 14:50:35 -07:00
Olcan 3b025883b6
fix json import warning (#291) 2025-05-08 14:14:09 -07:00
Miguel Solorio 5db1b7622a
Make ascii logo simpler (#288) 2025-05-08 13:46:41 -07:00
Olcan 06e5dfd538
minor comment fix (#290) 2025-05-08 11:31:12 -07:00
Olcan b59a940057
adjust seatbelt to allow write into specific dirs under user home (#289) 2025-05-08 11:28:45 -07:00
Tae Hyung Kim 448a24746c init 2025-05-07 23:47:58 -07:00
Taylor Mullen 6989032414 Remove unnecessary sleep.
- Code review comment: https://github.com/google-gemini/gemini-code/pull/271#pullrequestreview-2821741430
2025-05-07 23:46:57 -07:00
Olcan 327bd5f836
rename SANDBOX_EXEC_PROFILE as SEATBELT_PROFILE, and fix another accidental rephrasing (#285) 2025-05-07 21:31:30 -07:00
Tae Hyung Kim 13eadcea45
Fix bugs from useGeminiStream refactor (#284) 2025-05-07 21:15:41 -07:00
Olcan d524309e3c
use seatbelt on macos, with two profiles: minimal (default) which only restricts writes, and strict, which is deny-by-default and only allows specific operations (#283) 2025-05-07 20:03:29 -07:00
Taylor Mullen 34fe142894 Update EditTool description for clarity and better parameter guidance.
- Prior to this change, the model would often escape parameters when requesting edits, leading to failures in matching the original content. This update clarifies the expected format for `old_string` and `new_string` to prevent such issues.
   - Update `EditTool` description to provide clearer instructions.
   - Clarify expectations for `old_string` and `new_string` parameters, emphasizing the need for exact, unescaped text.
   - Aim to reduce user errors by setting better expectations for tool usage.

Fixes: https://b.corp.google.com/issues/413088274

— Your friendly neighborhood Gemini
2025-05-07 18:31:39 -07:00
Taylor Mullen 43c707b4e8 Continue to work through 429/500s.
- The root of this issue was actually a genai SDK bug that was fixed here: https://critique.corp.google.com/cl/753255997
- Upgrade to latest genai SDK for latest bug fixes (including the above)
- Removed specific 429 handling for uncaught rejections.

Fixes https://b.corp.google.com/issues/413760164
2025-05-07 16:38:06 -07:00