On 4/21/22 05:37, Zhiyi Zhang wrote:
+static HRESULT STDMETHODCALLTYPE api_information_statics_IsApiContractPresentByMajor( + IApiInformationStatics *iface, HSTRING contract_name, UINT16 major_version, BOOLEAN *value) +{ + FIXME("iface %p, contract_name %s, major_version %u, value %p stub!\n", iface, + debugstr_hstring(contract_name), major_version, value); + + if (value) + { + *value = FALSE; + return S_OK; + } + + return E_NOTIMPL; +}
I find it weird to have a mix of E_NOTIMPL / non-error result, especially when it only depend on the pointer validity. IMHO the patch should only stub the method, returning E_NOTIMPL, and then another patch with a small test should check whether value can be NULL, and semi-stub it to always return FALSE / S_OK. -- Rémi Bernon <rbernon(a)codeweavers.com>