Re: AW: AW: [lwip-users] ioctl() plus compiler warnings/errors

[leon . woestenberg]

Florian,

what happens if
#define ((u8_t)TF_FOO) 0x08U

is used?

Leon.




|---------+------------------------------------------------------>
|         |           "Zschocke, Florian" <address@hidden>  |
|         |           Sent by:                                   |
|         |           lwip-users-bounces+leon.woestenberg=axon.tv|
|         |           @nongnu.org                                |
|         |                                                      |
|         |                                                      |
|         |           18-02-2003 14:37                           |
|         |           Please respond to Mailing list for lwIP    |
|         |           users                                      |
|         |                                                      |
|---------+------------------------------------------------------>
  
>----------------------------------------------------------------------------------------------|
  |                                                                             
                 |
  |       To:       'Mailing list for lwIP users' <address@hidden>              
          |
  |       cc:                                                                   
                 |
  |       Subject:  AW: AW: [lwip-users] ioctl() plus compiler warnings/errors  
                 |
  
>----------------------------------------------------------------------------------------------|




address@hidden wrote on Dienstag, 18. Februar 2003 13:40:

> But what if you define 'c' as being a 'unsigned char'? If you
> are using it as a set of flags, this should really work.

Apparently not, since the tcp_pbc::flags is an u8_t. And that is where it
occurs.

tcp_out.c:377:     pcb->flags &= ~(TF_ACK_DELAY | TF_ACK_NOW);

results in:

"..\src\core\tcp_out.c", line 377: warning (dcc:1244): constant out of
range
(&=)

But since that may only a problem for us I'm fine when you decide not to
use
the patch. We still have the option to either keep our sources changed or
turn the compiler error into a warning and ignore it. But I would still
suggest to use unsigned long for the cmd parameter of the lwip_ioctl()
function.

Florian.


_______________________________________________
lwip-users mailing list
address@hidden
http://mail.nongnu.org/mailman/listinfo/lwip-users