Commit graph

1268 commits

Author SHA1 Message Date
sam hoang
b4094a6281 refactor by pr comment 2025-03-04 21:56:22 +07:00
sam hoang
167229fa73 Refactor checkExistKey to use centralized SECRET_KEYS array 2025-03-04 21:55:05 +07:00
sam hoang
381b07849a Feat ContextProxy to improve state management
- Add ContextProxy class as a wrapper around VSCode's ExtensionContext
- Implement batched state updates for performance optimization
- Update ClineProvider to use ContextProxy instead of direct context access
- Add comprehensive test coverage for ContextProxy
- Extract SECRET_KEYS and GLOBAL_STATE_KEYS constants for better maintainability
2025-03-04 21:55:04 +07:00
Matt Rubens
541b54e015
Merge pull request #1368 from adamwlarson/speculative_decoding
Adding support for Speculative Decoding for LMStudio Local Models
2025-03-04 09:32:58 -05:00
cte
d7fb9980e6 Revert logging in ClineProvider 2025-03-04 01:39:20 -08:00
cte
0bd56ad165 Checkpoints: one-branch-per-task / one-repo-per-workspace 2025-03-04 01:37:05 -08:00
Adam Larson
1c8f9ed683 first pass on speculative decoding with LMStudio 2025-03-04 01:48:40 -05:00
Matt Rubens
af06af136c
Set Vertex maxTokens to 64000 2025-03-03 16:03:22 -05:00
cte
e227a9e86e Remove unused method 2025-03-03 12:35:00 -08:00
cte
9e471444de Async checkpoints 2025-03-03 12:28:16 -08:00
Chris Estreich
464d440560
Merge pull request #1319 from samhvw8/feat/windown-path-and-escape-space-path
feat(mentions): improve path handling for Windows and escaped spaces
2025-03-03 08:56:35 -08:00
sam hoang
41a1c82fc5 add test from pr comment 2025-03-03 22:59:13 +07:00
Matt Rubens
003cd1bce1 Exclude MCP from system prompt if mode does not support it 2025-03-03 10:46:54 -05:00
Matt Rubens
2b7df5e106 Update max tokens for vertex too 2025-03-03 09:44:02 -05:00
Matt Rubens
63cecf8b22 Support 128k max tokens for openrouter thinking 2025-03-03 09:37:55 -05:00
Matt Rubens
6cf420235d
Merge pull request #1327 from monotykamary/feat/claude-extended-output
feat: update virtual id for extended output 128k claude
2025-03-03 09:29:28 -05:00
Matt Rubens
6b0347e0bd Add a checkbox to enable/disable the browser tool 2025-03-03 00:34:20 -05:00
Tom X Nguyen
ef93f5d422 refactor: always enable 128k for simplification 2025-03-03 11:17:01 +07:00
Tom X Nguyen
38cf054391 chore: add missing 3.7 sonnet modelid 2025-03-03 09:11:18 +07:00
Matt Rubens
9b91ab21b6 Default middle-out compression to on 2025-03-02 20:57:53 -05:00
Tom X Nguyen
d2b4a83605 fix: simplify logic and allow for multiple betas 2025-03-03 08:33:35 +07:00
Tom X Nguyen
aec03d71c3 feat: add virtual id for extended output capabilities 2025-03-03 07:47:12 +07:00
sam hoang
dd961d18a8 feat(mentions): improve path handling for Windows and escaped spaces
Enhance the mention regex to properly support:
- Windows paths with drive letters (C:\folder\file.txt)
- Windows network shares (\\server\share\file.txt)
- Windows relative paths (folder\file.txt)
- Paths with escaped spaces in both Unix and Windows formats
- Maintain compatibility with existing URL, git hash, and keyword patterns
- Add comprehensive test suite with 200+ test cases validating all path formats
and edge cases to ensure reliable pattern matching across platforms.
2025-03-02 23:56:35 +07:00
Matt Rubens
7e62d34570 PR cleanup 2025-03-02 02:12:21 -05:00
Matt Rubens
e8de7dea19 Merge remote-tracking branch 'origin/main' into count_tokens 2025-03-02 02:02:05 -05:00
Matt Rubens
8a1d66732a Infrastructure to support calling token count APIs, starting with Anthropic 2025-03-02 01:47:55 -05:00
cte
9b646eb2bc DRY up getModel 2025-03-01 21:51:17 -08:00
Matt Rubens
11387bdecd
Merge pull request #1304 from RooVetGit/support_read_only_modes_that_can_run_commands
Clearer instructions for read-only modes that can run commands
2025-03-02 00:15:19 -05:00
Matt Rubens
41b0ddecb4 Clearer instructions for read-only modes that can run commands 2025-03-02 00:13:14 -05:00
cte
62385ba4c7 Custom max tokens fix for non-thinking models 2025-03-01 21:05:03 -08:00
Matt Rubens
e6a49ddc5c Add type='module' to webview script 2025-03-01 20:42:10 -05:00
Matt Rubens
22c8a683d8
Merge pull request #1296 from RooVetGit/mermaids
Mermaids
2025-03-01 18:51:40 -05:00
Matt Rubens
8c8210ef50
Merge pull request #1282 from ashktn/feat/vertex-gemini
feat: Add support for Gemini models on Vertex AI
2025-03-01 18:48:54 -05:00
Matt Rubens
45eef8e37f Update snapshots 2025-03-01 18:46:35 -05:00
Matt Rubens
c8b72fa730 Update ask prompt 2025-03-01 18:19:16 -05:00
Matt Rubens
65a2a2cbd1 Update architect prompt 2025-03-01 18:10:33 -05:00
Matt Rubens
a0684454a2 Add a dynamic token buffer 2025-03-01 10:10:24 -05:00
Matt Rubens
764d963c10 Add a 5k token buffer before the end of the context window 2025-03-01 08:49:10 -05:00
Chris Estreich
481d613bc8 Estimate the number of tokens in the last message for sliding window math 2025-02-28 21:34:53 -08:00
Chris Estreich
1d3f5380a8 Move integration tests into its own module 2025-02-28 21:23:11 -08:00
ashktn
4caa49228f feat: Add support for Gemini models on Vertex AI 2025-02-28 18:59:26 -05:00
Chris Estreich
2773208d06
Merge pull request #1276 from RooVetGit/cte/re-enable-tests 2025-02-28 12:07:32 -08:00
Chris Estreich
55466c9d6a
Merge pull request #1268 from feifei325/optimize-task-id-init
Optimize Task ID Initialization Order
2025-02-28 11:15:07 -08:00
Chris Estreich
e49d886d63 Re-enable tests 2025-02-28 10:21:52 -08:00
Chris Estreich
21b883b59b
Merge pull request #1236 from RooVetGit/cte/fix-terminal-output-parsing
Fix TerminalProcess output parsing
2025-02-28 10:18:07 -08:00
Chris Estreich
4b66ce8255 Improve terminal command execution output parsing 2025-02-28 09:38:48 -08:00
yansheng3
0a81853d84 Optimize Task ID Initialization Order
Signed-off-by: yansheng3 <yansheng3@staff.weibo.com>
2025-02-28 19:28:45 +08:00
Chris Estreich
31d2d17847 Fix maxTokens defaults for Claude 3.7 Sonnet models 2025-02-28 01:15:44 -08:00
cte
8cbce2ded0 Add provider-agnostic modelMaxThinkingTokens setting 2025-02-27 16:06:47 -08:00
Chris Estreich
54c687485b
Merge pull request #1254 from lupuletic/added-new-thinking-support
Added support for Claude Sonnet 3.7 thinking via Vertex AI
2025-02-27 15:51:58 -08:00