ralf.habacker at freenet.de
Thu Feb 25 10:41:28 PST 2010
Am 25.02.2010 17:58, schrieb Colin Walters:
> On Wed, Feb 24, 2010 at 9:04 AM, Ralf Habacker<ralf.habacker at freenet.de> wrote:
> First, there's various trailing whitespace in the patch.
This may also come from windows/unix eol difference.
> You might
> find my script for
> fixing this useful:
Thanks I tried it with success. +dbus_bool_t
> +_dbus_string_tolower_ascii (const DBusString *str,
> + int start,
> + int len)
> Thinking about it I really don't like this function returning a bool
> for whether or not the passed in bounds were valid. Every single
> other function in dbus-string.h (and really, DBus in general) uses the
> boolean for out of memory.
> The range check should just be _dbus_assert I think, and this function
> should return void.
Is this really good ? As you mentioned in  there are many test cases
not implementable. In fact I'm only allowed to write test cases for oom
errors, which will not cover the complete range of possible function
>> +#if defined(_DEBUG)&& defined(DBUS_WIN)
>> + char *dummy1; /**< placeholder */
>> const void *dummy1; /**< placehlder */
> Also what is the rationale behind this change? It seems totally unrelated...
Otherwise it is very hard to see the content of a DBusString for
debugging purpose, but sure, I can may a different patch.
More information about the dbus