From: Alexandre Julliard <julliard@winehq.org>
Subject: Re: [PATCH 2/2] server: Strip terminating '\0' from the registry key name. (Resend)
Message-Id: <877daoyrgf.fsf@wine>
Date: Tue, 25 Jan 2022 12:39:44 +0100
In-Reply-To: <20220125141124.1284073c7ca000048266a0d7@baikal.ru> (Dmitry Timoshkov's message of "Tue, 25 Jan 2022 14:11:24 +0300")
References: <20220125115646.e681d67765dfa016f203e9e3@baikal.ru> <87czkgytwo.fsf@wine> <20220125141124.1284073c7ca000048266a0d7@baikal.ru>

Dmitry Timoshkov <dmitry@baikal.ru> writes:

> Alexandre Julliard <julliard@winehq.org> wrote:
>
>> > @@ -477,6 +477,8 @@ static inline void get_req_path( struct unicode_str *str )
>> >  {
>> >      str->str = get_req_data();
>> >      str->len = (get_req_data_size() / sizeof(WCHAR)) * sizeof(WCHAR);
>> > +
>> > +    while (str->len && !str->str[str->len / sizeof(WCHAR) - 1]) str->len -= sizeof(WCHAR);
>> 
>> I don't see a test that would justify that while() loop. You should also
>> test embedded nulls.
>
> Is that acceptable to test embedded nulls in a separate patch?

I'm not sure I see why, you need the test to know how to implement it.

-- 
Alexandre Julliard
julliard@winehq.org