| Re: [AD] Patches for allegro 4.1.9 |
[ Thread Index |
Date Index
| More lists.liballeg.org/allegro-developers Archives
]
* Eric Botcazou (ebotcazou@xxxxxxxxxx) [20030205 12:44]:
> Note that you can already specify -mcpu and -march options at configure time
> (see --enable-opts and --enable-exclopts).
Ok, I hadn't noticed these.
> The function is a method of the INTERNAL_MOUSE_DRIVER class, so we would need
> to modify every mouse driver.
The question is, if this wouldn't be the cleaner way to deal with it. But if
you don't want to go that route, there should at least be a check that the
passed in value isn't negative.
> I hate double-underscored types :-)
Well, C99 has uint32_t .... So you'd prefer to use 'unsigned int' here?
> I'm going be more drastic: Allegro has compiled for years with -Wall -W
> -Werror under DOS and Windows so I don't see why the Unix port should be
> less strict. Your patch is almost sufficient to achieve this goal (there are
> some incomplete initializers that I'm going to fill).
If you use -W -Wall -Werror, unused parameters have to be taken care of or
alternatively -Wno-unused-parameters be used. In case you don't object,
I could add an ATTRIBUTE_UNUSED to one of the headers that expands to
__attribute__((unused)) on gcc >= 2.7 and to nothing otherwise. In that case
all places where unused parameters occur, they would have to be marked with
ATTIBUTE_UNUSED.
Which way would you prefer?
Philipp
--
Philipp Thomas <pthomas@xxxxxxxxxx>
SuSE Linux AG, Deutschherrnstr. 15-19, D-90429 Nuremberg, Germany