On 09/02/2014 12:47 PM, Fabiano Fidêncio wrote:
Hi,

On Tue, 2014-09-02 at 12:33 +0300, Uri Lublin wrote:
On 09/01/2014 04:45 PM, Fabiano Fidêncio wrote:
---
   common/canvas_utils.c  |  3 ++-
   common/lines.c         | 19 ++++++++++++++-----
   common/marshaller.c    |  2 +-
   common/pixman_utils.c  |  2 +-
   common/quic.c          |  2 +-
   common/quic_rgb_tmpl.c |  8 ++++----
   common/quic_tmpl.c     |  8 ++++----
   common/rop3.c          | 14 +++++++++-----
   8 files changed, 36 insertions(+), 22 deletions(-)

diff --git a/common/pixman_utils.c b/common/pixman_utils.c
index db7a67f..5bf5717 100644
--- a/common/pixman_utils.c
+++ b/common/pixman_utils.c
@@ -30,7 +30,7 @@
#define SOLID_RASTER_OP(_name, _size, _type, _equation) \
   static void                                        \
-solid_rop_ ## _name ## _ ## _size (_type *ptr, int len, _type src)  \
+solid_rop_ ## _name ## _ ## _size (_type *ptr, int len, SPICE_GNUC_UNUSED 
_type src)  \
   {                                                  \
       while (len--) {                                \
           _type dst = *ptr;                          \

Hi,

Just wanted to mention that src is actually used within this function
for most OPs.
It is hidden within _equation.
For some operations (such as "clear" and "noop") src is not used.
I completely overlooked at this case, you're right (and GCC is a bit
dumb for catching this up).

I guess the patch is OK (but a bit confusing), as gcc documentation [1],
explains that attribute unused means "to be possibly unused"
[1] https://gcc.gnu.org/onlinedocs/gcc/Variable-Attributes.html
Yeah, I'll remove this change and resubmit this patch.

If you remove this change, you'll likely get some errors/warnings for those OPs that do not use src.
I'd keep the change and add a comment that for most OPs 'src' is used but
for some it is not used, so the attribute added makes gcc happy.

_______________________________________________
Spice-devel mailing list
Spice-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/spice-devel

Reply via email to