Releases: stoatworks-labs/abomerration
Release list
v0.1.7
Resolume's browser has no vendor grouping, so the Stoatworks plugins scatter through one alphabetical list and you have to remember every name to find them (cogwheel#26). Every plugin now reports itself with an SW prefix, so they sit together.
Abomerration→SW Abomerration
Nothing else moved. The four-character plugin ID a composition actually references is unchanged, and so are the bundle file names — an upgrade replaces what you have instead of leaving the old plugin beside the new one.
Full Changelog: v0.1.6...v0.1.7
v0.1.6
What's Changed
- build(deps): Bump actions/download-artifact from 4 to 8 by @dependabot[bot] in #5
- Keep a preset selected when the host restates its own values by @stoatworks-labs in #8
- Say what the plugin does, in the field Resolume shows by @stoatworks-labs in #9
Full Changelog: v0.1.5...v0.1.6
v0.1.5
v0.1.4
What's Changed
- build(deps): bump actions/upload-artifact from 4 to 7 by @dependabot[bot] in #1
- build(deps): bump actions/checkout from 4 to 7 by @dependabot[bot] in #2
- build(deps): Bump softprops/action-gh-release from 2 to 3 by @dependabot[bot] in #3
- Say what Edges does, in the sixteen characters FFGL allows by @stoatworks-labs in #6
- Release v0.1.4 by @stoatworks-labs in #7
New Contributors
- @dependabot[bot] made their first contribution in #1
- @stoatworks-labs made their first contribution in #6
Full Changelog: v0.1.3...v0.1.4
v0.1.3
The OpenFX plugin now ships for Linux
There is a -ofx-linux-x86_64 build in this release, for Resolve on Linux. It is
built against glibc 2.28 — the floor Rocky 8.6 provides, which is the Linux
Resolve supports — and checked by loading it in a Rocky 8 container and calling
the two entry points a host calls first, rather than inferring from a successful
build that it would load.
That test is not a formality. Four plugins in this fleet failed it the first time
with undefined symbol: pthread_create: at glibc 2.28 pthreads still live in
libpthread, and nothing linked it. They compiled, linked, exported OfxGetPlugin
and passed a glibc-version assertion first — a version check reads the symbols a
binary asks for, never whether anything provides them.
The FFGL plugin remains macOS and Windows only, because Resolume has no Linux
build to load it.
A clock check that had been failing, and why it was not the bug it looked like
abomtest --clock exists to catch one specific trap: Resolume sends SetTime
in milliseconds, the FFGL header never says so, and a plugin that consumes the
clock raw is a thousand times fast in the only host anybody runs it in — while
every other offline check still passes.
The check itself had been broken since the About-surface release. It declares the
clock unit to the plugin rather than making it infer one, because the inference
compares the host's clock against the wall clock and sixty frames render far
faster than real time. That declaration went into the code path shared by both
hosts and said "seconds" for each, so the millisecond host was being told its
milliseconds were seconds. The number it produced, 11.8 against a wanted 1.97,
was the per-frame clamp rather than a measurement — which is why it was six times
out instead of the thousand a genuine raw-clock bug gives.
The plugin was correct throughout; the harness was not. The fix declares the
unit per host, and the check is proven rather than assumed: making the plugin
ignore the unit puts the failure straight back.
All 23 checks in tools/verify.sh pass.
v0.1.2
v0.1.1
v0.1.0
Full Changelog: https://github.com/stoatworks-labs/abomerration/commits/v0.1.0