Skip to content

SCG + Ollama both (macOS26 and macOS15)#9

Open
manelix2000 wants to merge 3 commits intoIron-Ham:mainfrom
manelix2000:ollama
Open

SCG + Ollama both (macOS26 and macOS15)#9
manelix2000 wants to merge 3 commits intoIron-Ham:mainfrom
manelix2000:ollama

Conversation

@manelix2000
Copy link
Contributor

Fixed compilation errors on both macOS26 and macOS15. Now supports both operating systems in order to generate a version to use:

  • Sequoia (macOS15): use ollama as llm provider in order to use scg tool
  • Tahoe (macOS26): use both FoundationModels or ollama as llm provider in order to use scg tool. Select using scg config or as parameter in the command line execution

This commit updates key components of the SwiftCommitGen project, including improvements to the BatchCombinationPromptBuilder, LLMClient, and CommitGenError handling. These changes enhance the project's ability to manage LLM-based commit generation more robustly, ensuring better error handling and improved integration with FoundationModels. The updates also streamline the CommitGenTool's configuration process, making it more efficient for developers working with Ollama models. Additionally, the project now includes enhancements to DiffSummarizer.swift, CommitGenOptions.swift, and PromptBuilder.swift, adding new functionality and improving existing features. These modifications focus on enhancing the core components of the project, ensuring better performance and integration with newer macOS frameworks. This update is part of a larger batch of changes aimed at improving the project's capabilities and compatibility with modern development standards.
This commit updates several components in the SwiftCommitGen project, including enhancements to BatchCombinationPromptBuilder and LLMClient, and improvements to PromptBuilder. These changes aim to improve the project's functionality and maintainability. The modifications involve restructuring and refining code to ensure better performance and error handling.
The LLMClient struct has been refactored to reduce duplication of code and improve readability. The original implementation was using duplicate logic to handle different types of prompts, but this was causing confusion and difficult to maintain. By extracting the common logic into a separate function, we were able to simplify the code and make it easier to understand.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant