Re: [PULL 00/12] tcg patch queue

[Richard Henderson]

On 10/25/19 12:12 PM, Peter Maydell wrote:
> Compile failure, win32:
...
> /usr/lib/mxe/usr/x86_64-w64-mingw32.shared/include/glib-2.0/glib/gmacros.h:337:23:
> error: expected expression befor
> e 'do'
>  #define G_STMT_START  do
>                        ^
> /usr/lib/mxe/usr/x86_64-w64-mingw32.shared/include/glib-2.0/glib/gtestutils.h:115:41:
> note: in expansion of macro 'G_STMT_START'
>  #define g_assert(expr)                  G_STMT_START { \
>                                          ^
> /home/petmay01/qemu-for-merges/include/qemu/osdep.h:152:20: note: in
> expansion of macro 'g_assert'
>  #define assert(x)  g_assert(x)

Ho hum.   This...

>                     ^
> /home/petmay01/qemu-for-merges/include/exec/cpu-all.h:225:29: note: in
> expansion of macro 'assert'
>  #  define TARGET_PAGE_MASK (assert(target_page.decided), target_page.mask)

... must be the reason why we used to use ({ ... }) here.

Why oh why does g_assert use a do/while(0) statement instead of use an
expression like the C assert is required to do?

> Also
> /home/petmay01/qemu-for-merges/exec.c: In function 'iotlb_to_section':
> /home/petmay01/qemu-for-merges/exec.c:2871:1: error: control reaches
> end of non-void function [-Werror=return-type]
>  }
...
> Not sure if that's just follow-on errors from the earlier
> ones, though.

These are all follow-on, as you say.


r~