The Simple Authentication and Security Layer (SASL) library
re-defines the struct iovec on Win32 [*]. QEMU also re-defines
it in "qemu/osdep.h". The two definitions then clash on a MinGW
build.
We can avoid the SASL definition by defining STRUCT_IOVEC_DEFINED.
Add the definition to vnc_sasl_cflags if we are uing MinGW.

[*] https://github.com/cyrusimap/cyrus-sasl/blob/cyrus-sasl-2.1.27/include/sasl.h#L187

Cc: Alexey Pavlov <alexpux@gmail.com>
Cc: Biswapriyo Nath <nathbappai@gmail.com>
Cc: Youry Metlitsky <winaes@yandex.ru>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
---
 configure | 8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)

diff --git a/configure b/configure
index 0c2dd1eb08..0bc87ce42a 100755
--- a/configure
+++ b/configure
@@ -3375,7 +3375,13 @@ if test "$vnc" = "yes" && test "$vnc_sasl" != "no" ; then
 int main(void) { sasl_server_init(NULL, "qemu"); return 0; }
 EOF
   # Assuming Cyrus-SASL installed in /usr prefix
-  vnc_sasl_cflags=""
+  if test "$mingw32" = "yes" && test "$iovec" != "yes"; then
+    # QEMU defines struct iovec in "qemu/osdep.h",
+    # we don't want libsasl to redefine it in <sasl/sasl.h>.
+    vnc_sasl_cflags="-DSTRUCT_IOVEC_DEFINED"
+  else
+    vnc_sasl_cflags=""
+  fi
   vnc_sasl_libs="-lsasl2"
   if compile_prog "$vnc_sasl_cflags" "$vnc_sasl_libs" ; then
     vnc_sasl=yes
-- 
2.21.1

On Sat, Mar 07, 2020 at 06:22:21PM +0100, Philippe Mathieu-Daudé wrote:
> The Simple Authentication and Security Layer (SASL) library
> re-defines the struct iovec on Win32 [*]. QEMU also re-defines
> it in "qemu/osdep.h". The two definitions then clash on a MinGW
> build.
> We can avoid the SASL definition by defining STRUCT_IOVEC_DEFINED.
> Add the definition to vnc_sasl_cflags if we are uing MinGW.
> 
> [*] https://github.com/cyrusimap/cyrus-sasl/blob/cyrus-sasl-2.1.27/include/sasl.h#L187
> 
> Cc: Alexey Pavlov <alexpux@gmail.com>
> Cc: Biswapriyo Nath <nathbappai@gmail.com>
> Cc: Youry Metlitsky <winaes@yandex.ru>
> Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
> ---
>  configure | 8 +++++++-
>  1 file changed, 7 insertions(+), 1 deletion(-)
> 
> diff --git a/configure b/configure
> index 0c2dd1eb08..0bc87ce42a 100755
> --- a/configure
> +++ b/configure
> @@ -3375,7 +3375,13 @@ if test "$vnc" = "yes" && test "$vnc_sasl" != "no" ; then
>  int main(void) { sasl_server_init(NULL, "qemu"); return 0; }
>  EOF
>    # Assuming Cyrus-SASL installed in /usr prefix
> -  vnc_sasl_cflags=""
> +  if test "$mingw32" = "yes" && test "$iovec" != "yes"; then

I don't get why we need the "iovec != yes" check there ?

On mingw sys/uio.h doesn't exist, so "$iovec" will always
be "no", and so this conditional is equivalent to

   if test "$mingw32" = "yes; then


If for some strange reason, a future Windows adds sys/uio.h
header containing struct iovec, then QEMU won't define its
own local copy if struct iovec, and so "$iovec" will be "yes".

In this situation we don't want SASL to define its struct iovec 
either.

IOW we need -DSTRUCT_IOVEC_DEFINED no matter what $iovec value
is AFAICT.

This would mean the previous patch is redundant too.

> +    # QEMU defines struct iovec in "qemu/osdep.h",
> +    # we don't want libsasl to redefine it in <sasl/sasl.h>.
> +    vnc_sasl_cflags="-DSTRUCT_IOVEC_DEFINED"
> +  else
> +    vnc_sasl_cflags=""
> +  fi
>    vnc_sasl_libs="-lsasl2"
>    if compile_prog "$vnc_sasl_cflags" "$vnc_sasl_libs" ; then
>      vnc_sasl=yes

Regards,
Daniel
-- 
|: https://berrange.com      -o-    https://www.flickr.com/photos/dberrange :|
|: https://libvirt.org         -o-            https://fstop138.berrange.com :|
|: https://entangle-photo.org    -o-    https://www.instagram.com/dberrange :|

On 3/9/20 10:57 AM, Daniel P. Berrangé wrote:
> On Sat, Mar 07, 2020 at 06:22:21PM +0100, Philippe Mathieu-Daudé wrote:
>> The Simple Authentication and Security Layer (SASL) library
>> re-defines the struct iovec on Win32 [*]. QEMU also re-defines
>> it in "qemu/osdep.h". The two definitions then clash on a MinGW
>> build.
>> We can avoid the SASL definition by defining STRUCT_IOVEC_DEFINED.
>> Add the definition to vnc_sasl_cflags if we are uing MinGW.
>>
>> [*] https://github.com/cyrusimap/cyrus-sasl/blob/cyrus-sasl-2.1.27/include/sasl.h#L187
>>
>> Cc: Alexey Pavlov <alexpux@gmail.com>
>> Cc: Biswapriyo Nath <nathbappai@gmail.com>
>> Cc: Youry Metlitsky <winaes@yandex.ru>
>> Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
>> ---
>>   configure | 8 +++++++-
>>   1 file changed, 7 insertions(+), 1 deletion(-)
>>
>> diff --git a/configure b/configure
>> index 0c2dd1eb08..0bc87ce42a 100755
>> --- a/configure
>> +++ b/configure
>> @@ -3375,7 +3375,13 @@ if test "$vnc" = "yes" && test "$vnc_sasl" != "no" ; then
>>   int main(void) { sasl_server_init(NULL, "qemu"); return 0; }
>>   EOF
>>     # Assuming Cyrus-SASL installed in /usr prefix
>> -  vnc_sasl_cflags=""
>> +  if test "$mingw32" = "yes" && test "$iovec" != "yes"; then
> 
> I don't get why we need the "iovec != yes" check there ?
> 
> On mingw sys/uio.h doesn't exist, so "$iovec" will always
> be "no", and so this conditional is equivalent to
> 
>     if test "$mingw32" = "yes; then

Indeed.

> 
> If for some strange reason, a future Windows adds sys/uio.h
> header containing struct iovec, then QEMU won't define its
> own local copy if struct iovec, and so "$iovec" will be "yes".
> 
> In this situation we don't want SASL to define its struct iovec
> either.
> 
> IOW we need -DSTRUCT_IOVEC_DEFINED no matter what $iovec value
> is AFAICT.

Way simpler!

> 
> This would mean the previous patch is redundant too.

Yes, thanks for reviewing :)

> 
>> +    # QEMU defines struct iovec in "qemu/osdep.h",
>> +    # we don't want libsasl to redefine it in <sasl/sasl.h>.
>> +    vnc_sasl_cflags="-DSTRUCT_IOVEC_DEFINED"
>> +  else
>> +    vnc_sasl_cflags=""
>> +  fi
>>     vnc_sasl_libs="-lsasl2"
>>     if compile_prog "$vnc_sasl_cflags" "$vnc_sasl_libs" ; then
>>       vnc_sasl=yes
> 
> Regards,
> Daniel
>