Re: [PATCH] advapi32/tests: Time zones should have both MUI and regular strings.

[Giovanni Mascellani]

Hi,

Il 17/05/21 14:41, Francois Gouget ha scritto:
> On Mon, 17 May 2021, Giovanni Mascellani wrote:
> [...]
> > > -
> > >            /* recently added time zones may not have MUI strings */
> > > -
> > >            todo_wine ok((status == ERROR_SUCCESS && *name) ||
> > > -
> > >               broken(status == ERROR_RESOURCE_TYPE_NOT_FOUND) /* Win10
> > >
> > > 1809 32-bit */ ||
> > > -
> > >               broken(status == ERROR_MUI_FILE_NOT_FOUND) /* Win10 1809
> > >
> > > 64-bit */,
> > > -
> > >               "status %d MUI_Display %s\n", status,
> > >
> > > wine_dbgstr_w(name));
> >
> > Why don't you check for *name after the first two calls to pRegLoadMUIStringW?
>
> It would be redundant with the memcmp() that follows?
> (or I misunderstood)
>              status = pRegLoadMUIStringW(subkey, L"MUI_Std", name, size, &size, 0, sysdir);
>              ok(status == ERROR_SUCCESS, "status %d MUI_Std %s\n",
>                 status, wine_dbgstr_w(name));
>          }
>          ok(!memcmp(&dtzi.StandardName, name, size), "expected %s, got %s\n",
>             wine_dbgstr_w(name), wine_dbgstr_w(dtzi.StandardName));

Unless I am mistaken, memcmp() checks that dtzi.StandardName and name 
are the same string (or, technically, that name is a prefix of 
dtzi.StandardName, which might be longer). They could be both empty strings.

BTW, this is not really a big problem in practice. I asked just because 
I saw that in the case of Display you did this check.

Thanks, Giovanni.