### Description
This PR fixes the reconnect / limited-sync recovery path in the room
timeline. Managed to track it down, thanks to this sentry report
[SABLE-1V](https://sable-client.sentry.io/issues/104545427/events/741b82918ba34554996ad673efe75a6f/)
After `TimelineReset`, the hook now re-anchors live event detection to
the current live timeline, allows bottom-pinned recovery to bypass the
transient `liveTimelineLinked` lag, and avoids snapping non-bottom users
to latest.
It also adds a regression test covering the reset behavior so the scroll
recovery and non-bottom path stay locked down.
Fixes#360
Repro (It's a tricky one):
- Use a broken build on Android Chrome if possible. Desktop Chrome works
too, but mobile/backgrounding is closer to the original report.
- Open a busy room and stay pinned to the bottom.
- In another client/account, send several messages to that room while
the first client is disconnected.
- Force a real sync gap:
- Mobile: background the app/tab, lock the phone, or toggle Wi‑Fi/mobile
data off for 30-60s.
- Desktop: open DevTools and set network to Offline for 30-60s.
- Reconnect and return to the room.
#### Type of change
- [x] Bug fix (non-breaking change which fixes an issue)
- [ ] New feature (non-breaking change which adds functionality)
- [ ] Breaking change (fix or feature that would cause existing
functionality to not work as expected)
- [ ] This change requires a documentation update
### Checklist:
- [x] My code follows the style guidelines of this project
- [x] I have performed a self-review of my own code
- [x] I have commented my code, particularly in hard-to-understand areas
- [ ] I have made corresponding changes to the documentation
- [x] My changes generate no new warnings
### AI disclosure:
- [ ] Partially AI assisted (clarify which code was AI assisted and
briefly explain what it does).
- [x] Fully AI generated (explain what all the generated code does in
moderate detail).
<!-- Write any explanation required here, but do not generate the
explanation using AI!! You must prove you understand what the code in
this PR does. -->
<!-- Please read
https://github.com/ajbura/cinny/blob/dev/CONTRIBUTING.md before
submitting your pull request -->
### Description
When using the touchpad both browser and app zoom gets activated, so let
the browser handle zooming when using a touchpad
Fixes#463
#### Type of change
- [x] Bug fix (non-breaking change which fixes an issue)
- [ ] New feature (non-breaking change which adds functionality)
- [ ] Breaking change (fix or feature that would cause existing
functionality to not work as expected)
- [ ] This change requires a documentation update
### Checklist:
- [x] My code follows the style guidelines of this project
- [x] I have performed a self-review of my own code
- [x] I have commented my code, particularly in hard-to-understand areas
- [ ] I have made corresponding changes to the documentation
- [x] My changes generate no new warnings
### Description
Fixes#285 for good. Element Call (and thus Sable Call) use a
`callIntent` paramter derived from the `intent` parameter to decide if
the Camera should be requested. The Camera was being requested on every
call, and thus auto approved on Electron clients (and Tauri clients).
This affected both starting a room call and joining a VC room. This is
likely due to the fact that in Element they are 'Video Rooms'. This
change simply uses `JoinExistingDMVoice` which sets the correct
parameters to join a VC room OR call from the room header. The
`StartCall` and `JoinExisting` intents function the same on Sable Call's
end, this can be changed if they start to diverge in future updates of
course though.
I have tested this change in both scenarios of calling with multiple
people and it has functioned as intended.
#### Type of change
- [x] Bug fix (non-breaking change which fixes an issue)
- [ ] New feature (non-breaking change which adds functionality)
- [ ] Breaking change (fix or feature that would cause existing
functionality to not work as expected)
- [ ] This change requires a documentation update
### Checklist:
- [x] My code follows the style guidelines of this project
- [x] I have performed a self-review of my own code
- [ ] I have commented my code, particularly in hard-to-understand areas
- [ ] I have made corresponding changes to the documentation
- [x] My changes generate no new warnings
### AI disclosure:
- [ ] Partially AI assisted (clarify which code was AI assisted and
briefly explain what it does).
- [ ] Fully AI generated (explain what all the generated code does in
moderate detail).
<!-- Write any explanation required here, but do not generate the
explanation using AI!! You must prove you understand what the code in
this PR does. -->
### Description
This PR adds the requirement to disclose AI usage on the PR, asking for
an explanation of what any AI generated code does.
With this I hope the AI dramas can end, while also making sure that no
vibecoded bad PR gets through, only the good ones.
I'm open to discussion but let's keep it calm here...
(Oh also, this PR fixes the link at the top of the PR template to link
to our contributing page and not Cinny's.)
#### Type of change
- [ ] Bug fix (non-breaking change which fixes an issue)
- [ ] New feature (non-breaking change which adds functionality)
- [ ] Breaking change (fix or feature that would cause existing
functionality to not work as expected)
- [ ] This change requires a documentation update
### Checklist:
- [ ] My code follows the style guidelines of this project
- [ ] I have performed a self-review of my own code
- [ ] I have commented my code, particularly in hard-to-understand areas
- [ ] I have made corresponding changes to the documentation
- [x] My changes generate no new warnings
### Description
Group npm Dependabot version updates into separate major and non-major
PR buckets so routine updates stay consolidated while breaking changes
remain isolated.
Fixes #
#### Type of change
- [x] Bug fix (non-breaking change which fixes an issue)
- [ ] New feature (non-breaking change which adds functionality)
- [ ] Breaking change (fix or feature that would cause existing
functionality to not work as expected)
- [ ] This change requires a documentation update
### Checklist:
- [x] My code follows the style guidelines of this project
- [x] I have performed a self-review of my own code
- [ ] I have commented my code, particularly in hard-to-understand areas
- [ ] I have made corresponding changes to the documentation
- [x] My changes generate no new warnings
### Description
I've added maskable icons to the selection of app icons. Maskable icons
are designed to be cropped into any shape, like on various Android
launchers, and this distinction additionally makes them work well on
iOS.
Here is a screenshot for comparison, with the new icon on the left :
<img width="522" height="323" alt="image"
src="https://github.com/user-attachments/assets/a647c9a3-c2cf-4f8a-8fee-5a5c15829be5"
/>
In the process of doing this, I've adjusted the structure of the
directories containing the app icons (they are no longer sorted by
Android/Apple, since any platform can pull from either format under
different circumstances). On that note, it looks like many of the
resolutions could probably be pruned... I kept all of the existing ones,
though.
I've also removed the dots from the paths in the HTML file, since this
was causing major issues in the `dev` run. I haven't been able to do a
proper build and test by hosting the /dist folder, so if this does break
anything in that scenario, I can revert that change.
I've tested this to completion on Linux and iOS so far, and I will edit
this part as I test more platforms (assuming it isn't merged by then).
- [ ] Android: Seems to work; a bit stretched, though? Edge might just
be bad at making PWAs. Should try more browsers.
- [ ] Windows:
- [ ] macOS:
- [ ] iPadOS:
#### Type of change
- [ ] Bug fix (non-breaking change which fixes an issue)
- [x] New feature (non-breaking change which adds functionality)
- [ ] Breaking change (fix or feature that would cause existing
functionality to not work as expected)
- [ ] This change requires a documentation update
### Checklist:
- [x] My code follows the style guidelines of this project
- [x] I have performed a self-review of my own code
- [ ] I have commented my code, particularly in hard-to-understand areas
- [ ] I have made corresponding changes to the documentation
- [x] My changes generate no new warnings
- flatted >=3.4.2: fixes CVE-2026-33228 (prototype pollution via parse())
- undici >=7.24.0: fixes 6 CVEs including DoS via WebSocket frame overflow,
decompression bombs, request smuggling, and CRLF injection
Both affect dev dependencies only (vitest, jsdom, wrangler).
<!-- Please read
https://github.com/ajbura/cinny/blob/dev/CONTRIBUTING.md before
submitting your pull request -->
### Description
This adds x-m4a to the supported file type list, this is the same mime
type that element uses for m4a files making Sable compatible with both
uploading them and playing them.
Fixes #
#### Type of change
- [X] Bug fix (non-breaking change which fixes an issue)
- [ ] New feature (non-breaking change which adds functionality)
- [ ] Breaking change (fix or feature that would cause existing
functionality to not work as expected)
- [ ] This change requires a documentation update
### Checklist:
- [X] My code follows the style guidelines of this project
- [X] I have performed a self-review of my own code
- [ ] I have commented my code, particularly in hard-to-understand areas
(Would have been overkill for what it is)
- [ ] I have made corresponding changes to the documentation (Is this
needed for this simple PR?)
- [X] My changes generate no new warnings
Prevents layout shift by rendering the recorder in the editor's main content
area instead of in the after buttons section. Uses replacementContent prop
to swap the editor with the recorder while maintaining proper padding.
When recording on mobile, if the pointerup listener was lost (e.g., dragging
off the button), there was no way to stop recording. Now the stop button's
onClick handler works on mobile when recording is active, providing a fallback.