Re: [PATCH v3 0/2] MR2231: cfgmgr32: Implement CM_MapCrToWin32Err.

[Zebediah Figura (＠zfigura)]

Zebediah Figura (@zfigura) commented about dlls/cfgmgr32/tests/cfgmgr32.c:
> -
>    { CR_QUERY_VETOED,             DEFAULT_ERROR },
> -
>    { CR_CANT_SHARE_IRQ,           DEFAULT_ERROR },
> -
>    { CR_NO_DEPENDENT,             DEFAULT_ERROR },
> -
>    { CR_SAME_RESOURCES,           DEFAULT_ERROR },
> -
>    { CR_DEVICE_INTERFACE_ACTIVE,  DEFAULT_ERROR },
> -
>    { CR_INVALID_CONFLICT_LIST,    DEFAULT_ERROR },
> -
>    { CR_INVALID_INDEX,            DEFAULT_ERROR },
> -
>    { CR_INVALID_STRUCTURE_SIZE,   DEFAULT_ERROR },
> -
>    { NUM_CR_RESULTS,              DEFAULT_ERROR },
> - };
> -
> - SetLastError(0xdeadbeef);
> - ret = 0;
> - ret = CM_MapCrToWin32Err( 0xdeadbeef, DEFAULT_ERROR );
> - ok( ret == DEFAULT_ERROR, "expected default error, got %ld.\n", ret );
> - ok( GetLastError() == 0xdeadbeef, "unexpected error %ld.\n", GetLastError() );

The fact that DEFAULT_ERROR is 0xdeadbeef makes the GetLastError() test less interesting.

The fact that DEFAULT_ERROR is a macro makes this above observation less obvious, which is one reason I would just write it out everywhere.

(Also, there's a dead assignment to "ret" in here.)