Commit Graph
2 Commits
Author SHA1 Message Date
Dario LipicarandClaude Opus 5 9d247d7b72 docs: correct two stale api-style comments in the context test (#36)
Both claimed a std path that does not exist:

flake.nix said "the LIDL sidecar shipped with basicCpp.headers-std
drives the codegen". Measured wrong twice over. test_basic_module_cpp
is `interface: universal`, so it publishes a `lidl` output and is
routed through `--dep test_basic_module_cpp=<its .lidl>` — it never
reaches the header-copy path, and its plugin is not built for
contextCpp at all. It is `test_basic_module` (no `interface` key, so no
`lidl`) that takes the header copy, and contextCpp is lp-typed, so it
copies that dep's `headers-lp`. Confirmed against the built plugin's
direct inputs: headers-lp for test_basic_module, -lidl for
test_basic_module_cpp, no headers-std anywhere.

test_context_module_cpp_impl.cpp said mkLogosModule passes
`-DLOGOS_API_STYLE=std` for universal modules. It passes lp (for core
universal modules — ui_qml view backends stay Qt-typed).

Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
2026-07-30 21:44:10 -03:00
Dario Lipicar 82cba02168 test universal modules with LogosModuleContext (#19)
* test universal modules with LogosModuleContext

* pr comments

* bump dependencies
2026-05-19 16:07:52 -03:00