diff options
| author | Danilo M. <danix@danix.xyz> | 2026-09-19 10:44:48 +0200 |
|---|---|---|
| committer | Danilo M. <danix@danix.xyz> | 2026-09-19 10:44:48 +0200 |
| commit | c10c750c83a3cde13f7ca574dac673e2358df522 (patch) | |
| tree | 11d3694da330091155e9c066399ce39c790000fc /CHANGELOG.md | |
| parent | 309c0dd8b0fc294f67af8a82b78830a7610b700f (diff) | |
| download | llamachat-c10c750c83a3cde13f7ca574dac673e2358df522.tar.gz llamachat-c10c750c83a3cde13f7ca574dac673e2358df522.zip | |
Make the chat toolbar icon-only (tooltips retained) and add an icon_size config key, default 28, so the drawn glyphs can be scaled. Qt was drawing the 24px pixmaps at its ~16px default; the explicit icon size fixes that.
Fix the deferred audio minors: AudioRecorder.start() now halts a prior source and checks the source error, so a failed capture no longer claims to be recording. Send now validates every pending modality in one gate, so a mixed image+audio clip cannot reach a model without vision. Sub-second clips no longer name voice-note-0s.wav, and non-WAV audio containers classify as unknown, guarded by a test.
Diffstat (limited to 'CHANGELOG.md')
| -rw-r--r-- | CHANGELOG.md | 18 |
1 files changed, 14 insertions, 4 deletions
diff --git a/CHANGELOG.md b/CHANGELOG.md index a2a3635..c685d54 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -69,10 +69,10 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0 moment once used. The text comes from the source markdown, since Qt emits one `<pre>` per line and the block cannot be recovered from the rendered HTML. -- Icons on the chat buttons (New, Attach, Clear files, Stop, Send), drawn - at runtime so there is still no icon file to install. Send is now a square - button beside the entry box with its label under the icon, and New/Attach - sit at the right end of the picker row. +- Icons on the chat buttons (New, Attach, Clear files, Stop, Send, Record), + drawn at runtime so there is still no icon file to install. The toolbar + buttons are icon-only, each with a tooltip, and `icon_size` in config.toml + sets the glyph size. - Skills. Instruction files read from `~/.agents/skills/*/SKILL.md` can be loaded into the model's context on demand: the model calls a `load_skill` tool, or you type `/skill-name` or mention a skill's name. Loaded skills @@ -118,6 +118,16 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0 assistant tool-call message. Interleaved-thinking models (DeepSeek V3.2+/V4, GLM-4.7+ on SiliconFlow) require this and stop answering — ending the turn at the thinking — when it is dropped. +- Audio capture no longer claims to be recording when it is not. + `AudioRecorder.start()` halts any previous source, checks the new source's + error, and reports a generic failure; a second start without a stop no + longer leaks the first source. +- A pending image attached before a model switch could reach a model without + vision when sent alongside audio. The send-time gate now checks every + pending modality at once and offers a single switch to a model that accepts + them all. +- A sub-second recording could be named `voice-note-0s.wav`; the history + marker now rounds to a minimum of one second. Tool calling and reasoning output vary between providers. Web search on a cloud model may not work as reliably as it does with a local router. |