The gtk4 host-drivers stage (experiments/50-gtk4-host-drivers.sh, added with E91/E92 and the GLFWD_ALT_SONAME repair) is reachable from both orchestrators but is missing from the documentation that lists the stages and the accepted values.
scripts/run-appimage.sh:41 accepts gtk4hd, and experiments/appimage.ps1:46 has it in its ValidateSet. Neither the workflow_dispatch input in .github/workflows/appimage-suite.yml:38 ("alpine, debian, ubuntu1404, ubuntu1604, gtk4, both, all") nor the list in docs/reproducing.md:58 mentions it.docs/reproducing.md:55 says "Five stages, not two" and describes five. With gtk4 and gtk4hd both present there are six (alpine, debian, ubuntu1404, ubuntu1604, gtk4, gtk4hd).scripts/run-appimage.sh:6 says "across four host classes and a second AppImage" and its header lists one gtk4 AppImage, not the second shape.
Fix: add gtk4hd to the workflow input and the docs/reproducing.md list, and correct the stage counts.
The gtk4 host-drivers stage (
experiments/50-gtk4-host-drivers.sh, added with E91/E92 and theGLFWD_ALT_SONAMErepair) is reachable from both orchestrators but is missing from the documentation that lists the stages and the accepted values.scripts/run-appimage.sh:41acceptsgtk4hd, andexperiments/appimage.ps1:46has it in itsValidateSet. Neither theworkflow_dispatchinput in.github/workflows/appimage-suite.yml:38("alpine, debian, ubuntu1404, ubuntu1604, gtk4, both, all") nor the list indocs/reproducing.md:58mentions it.docs/reproducing.md:55says "Five stages, not two" and describes five. Withgtk4andgtk4hdboth present there are six (alpine, debian, ubuntu1404, ubuntu1604, gtk4, gtk4hd).scripts/run-appimage.sh:6says "across four host classes and a second AppImage" and its header lists one gtk4 AppImage, not the second shape.Fix: add
gtk4hdto the workflow input and thedocs/reproducing.mdlist, and correct the stage counts.