Version 1.33.0 of package Ellama has just been released in GNU ELPA. You can now find it in M-x list-packages RET.
Ellama describes itself as: ============================== Tool for interacting with LLMs ============================== More at https://elpa.gnu.org/packages/ellama.html ## Summary: ━━━━━━━━ ELLAMA ━━━━━━━━ [file:https://img.shields.io/badge/license-GPL_3-green.svg] [file:https://melpa.org/packages/ellama-badge.svg] [file:https://stable.melpa.org/packages/ellama-badge.svg] [file:https://elpa.gnu.org/packages/ellama.svg] Ellama brings large language models into Emacs without turning Emacs into a web app. You can use it for ordinary chat, one-shot commands over the current buffer or region, and longer sessions that keep their history. It also has agent loops for work that needs several steps: make a plan, call tools, read project context, and continue from the previous turn. ## Recent NEWS: 1 Version 1.33.0 ════════════════ • Add `session_file' tool to expose the current session transcript as a file, allowing history-aware subagents to read and recover past session context during execution. • Omit tool results from session compaction by default. A new configurable opt-in (controlled by a dedicated custom) allows including tool results while keeping tool calls and arguments in the compaction input, keeping the compaction summary lean by default. • Integrate compacted tool-result recovery with the session transcript: tool calls and arguments are kept in the compaction summary so subagents can recover from compacted sessions, and transcript reads continue to obey session result truncation (SRT) policy. • Fix the compacted tool-result recovery assertion to match the new compaction behavior and add corresponding regression tests. • Fix the compaction hint docstring width so hints display correctly in the UI. • Extend `ellama-eval--instrument-tool-spec' to trace asynchronous tools whose first argument is a callback, and add a `balanced-edit-hooks' hypothesis profile that validates edits through a shell byte-compilation hook, surfaces hook failures to the agent, and records edit-hook failure and recovery counts in the evaluation summary. • Add `ellama-eval' configuration for lexical-binding declarations in eval fixture workspaces so all test fixtures run in lexical binding mode. • Add configuration to pin manual generation to Emacs 30.1 in CI, and let CI regenerate the committed manual on each build. 2 Version 1.32.3 ════════════════ • Use `start-file-process' for tool subprocesses so commands can run in remote TRAMP directories. 3 Version 1.32.2 ════════════════ • Add Emacs 31.1 to the MELPA and ERT CI matrices so new compiler warnings are caught in GitHub Actions. • Fix Emacs 31 compilation warnings by replacing deprecated conditional binding macros, providing a compatibility fallback for `replace-region-contents' on older Emacs versions, loading Info for the transient command, and capitalizing two warning messages to satisfy checkdoc. • Run native compilation in the Emacs 31.1 CI job and reformat Elisp code so the pre-push check passes. • Select buffer replacement by Emacs major version: older variants of `replace-region-contents' do not accept a buffer source argument. Retain Emacs 31.1 byte-compilation coverage without a native step, since the action build does not support it yet. 4 Version 1.32.1 ════════════════ • Preserve file buffer positions during tool updates. File contents are now updated with marker-preserving replacement, and only stale unmodified buffers are refreshed after shell commands and edit hooks. The system warns instead of discarding unsaved changes, and regression coverage for mutating, read-only, validation, and hook paths has been added. • Keep visiting buffers attached after file moves. The `move_file' tool now retargets an existing source buffer to the destination while preserving point and unsaved changes. Destinations already visited by another buffer are rejected. Regression tests cover both paths. 5 Version 1.32.0 ════════════════ • Preserve Markdown tables during streaming conversion. Paragraph filling no longer wraps table rows, and streaming fill skips while point is inside a table, avoiding phantom rows and broken cells. • Align converted Markdown tables in Org output. Separator rows are normalized and tables are aligned for both builtin and Pandoc converters, with better handling of wider cells during streaming updates. 6 Version 1.31.0 ════════════════ • Show irreversible action confirmation details. Named tool arguments are now included in security confirmation prompts, and a full-detail view flow has been added for DLP and irreversible actions. The flow preserves context even when audit logging fails, and regression tests cover prompt contents as well as the view interaction. 7 Version 1.30.0 ════════════════ • Split the subagent task tool into two separate contracts. The `task' tool handles free-form delegation, while `task_from_template' handles … …
