UPSTREAM PR #17976: common : add llama-completion to completion-bash executables

[loci-dev]

Mirrored from ggml-org/llama.cpp#17976

[loci-review]

Explore the complete analysis inside the Version Insights

Pull Request Review Summary: PR #541

Change Scope: Single-line addition to bash completion configuration
Files Modified: 1 (common/arg.cpp)
Performance Impact: None

Summary

This PR adds the llama-completion executable to the bash completion list in common_params_print_completion(). The change is purely a configuration update that enables shell autocompletion for the llama-completion tool. No performance-critical functions are modified. All core inference paths (llama_decode, llama_encode, llama_tokenize), tensor operations, and memory management functions remain unchanged. Power consumption analysis shows 0.0% change across all binaries. No impact on tokens per second.

[loci-review]

Explore the complete analysis inside the Version Insights

Pull Request Review Summary: PR #541

Change Scope: Single-line addition to bash completion configuration
Files Modified: 1 (common/arg.cpp)
Performance Impact: None

Summary

This PR adds the llama-completion executable to the bash completion list in common_params_print_completion(). The change is purely a configuration update that enables shell autocompletion for the llama-completion tool. No performance-critical functions are modified. All core inference paths (llama_decode, llama_encode, llama_tokenize), tensor operations, and memory management functions remain unchanged. Power consumption analysis shows 0.0% change across all binaries. No impact on tokens per second.

[loci-review]

Explore the complete analysis inside the Version Insights

Pull Request Review Summary: PR #541

Change Scope: Single-line addition to bash completion configuration
Files Modified: 1 (common/arg.cpp)
Performance Impact: None

Summary

This PR adds the llama-completion executable to the bash completion list in common_params_print_completion(). The change is purely a configuration update that enables shell autocompletion for the llama-completion tool. No performance-critical functions are modified. All core inference paths (llama_decode, llama_encode, llama_tokenize), tensor operations, and memory management functions remain unchanged. Power consumption analysis shows 0.0% change across all binaries. No impact on tokens per second.