5 Commits

Include renames

Author	SHA1	Message	Date
devsyncwrldandClaude Opus 5	293378293d	fix(ci): match release assets by name, and close the guard block ... Two defects, both mine, and the second hid the first. The cleanup built a regex and matched it against the jq line '<id> <name>', so the anchor in latest.yml could never match — it sits after the id. Only the .exe was removed, and each publish left another latest.yml behind. Gitea then served the older one, so the updater kept reporting the previous version as current: an update that exists and is never offered, with nothing logged anywhere. Filtering now happens inside jq, comparing the name directly. The guard meant to catch exactly that was missing its closing fi, so the step died on a syntax error before reaching it. Verified with bash -n this time, which is what should have happened before it ever ran on a runner. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>	2026-09-01 19:14:59 -03:00
devsyncwrldandClaude Opus 5	1fb61377b9	feat(notify): in-app notifications with the app's own sound ... The system balloon carries the OS notification sound, which does not belong to this app, and it only fired while the window was out of focus — with the app focused nothing appeared at all. Notifications now surface inside the window, carry the same synthesised timbre as the rest of the app's sounds, and clicking one opens the channel. The native balloon is kept for when the window is not visible, since an in-app card nobody can see is no notification, but it is now silent: the app plays its own effect instead. A focused window is notified only about other channels — announcing the conversation someone is already reading is noise. Also fixes the Gitea publish cleanup, which silently deleted nothing: it interpolated an Actions expression inside a bash , and when the pattern did not match, the loop passed over every asset. The release ended with two latest.yml files and the updater served the older one, reporting 1.1.0 as current — an update that exists but is never offered, with no error anywhere. The filter is plain bash now, logs what it found, and the job fails if more than one latest.yml survives. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>	2026-09-01 18:57:02 -03:00
devsyncwrld	ce6bba0510	fix(ci): pulseaudio headers, and an ASCII release payload ... Linux got past pipewire and stopped on pulse/pulseaudio.h — the module compiles both backends, so both sets of headers are needed. Windows reached the upload and Gitea rejected the release creation with HTTP 422, 'invalid UTF-8 within /name': the JSON body carried accented text and the Windows runner's shell mangles the encoding on the way out. The payload is now ASCII, with a comment saying why, so it does not get 'improved' back into Portuguese.	2026-09-01 16:04:01 -03:00
devsyncwrld	9ea399ded5	fix(ci): install pipewire headers, and make the Gitea upload diagnosable ... Two separate failures on the first run. Linux could not compile the native module: it needs libpipewire-0.3-dev, which its README states and the apt list omitted. Windows built everything, native module verified, then died on the upload with a bare JSONDecodeError. The cause was curl -sf, which discards the error body, so a failed release creation surfaced as an empty pipe and no reason at all. The step now captures status and body and prints them, treats a failed create as possibly the other matrix job having just created it, and passes target_commitish so the tag can be created. Embedded python gave way to jq: a multi-line heredoc inside a YAML literal block ends the block at the first unindented line, which is how the file became invalid YAML in the first place.	2026-09-01 15:58:26 -03:00
devsyncwrld	9f7723d104	ci: build on GitHub, publish updates from Gitea ... GitHub stays the build machine — it has the Windows runners the native audio module needs — but the update feed moves to this fork's own Gitea. The GitHub repository is private, and electron-updater against a private GitHub repo needs a token inside the shipped app, which is a leaked token. Gitea serves release assets to anyone, so the installer carries no credential. The flow was verified end to end against the live instance before writing this: create release, upload asset, download anonymously. The release tag is fixed at 'latest' because electron-updater fetches latest.yml before it knows which version exists, so the URL cannot carry a version; CI replaces that release's assets each publish. electron-builder runs with --publish never since it cannot upload to Gitea, but still emits the latest.yml the updater reads. Needs a GITEA_TOKEN secret on the GitHub repository.	2026-09-01 15:49:25 -03:00