Re: [PATCH 5/6] tests: handling signal on win32 properly

[Paolo Bonzini]

Yes, I guess that's true. I will queue it instead.

Paolo

Il gio 3 set 2020, 07:58 Thomas Huth <thuth@redhat.com> ha scritto:

On 02/09/2020 19.04, Paolo Bonzini wrote:
> On 02/09/20 19:00, Yonggang Luo wrote:
>> SIGABRT should use signal(SIGABRT, sigabrt_handler) to handle on win32
>>
>> The error:
>> E:/CI-Cor-Ready/xemu/qemu.org/tests/test-replication.c:559:33: error: invalid use of undefined type 'struct sigaction'
>>   559 |     sigact = (struct sigaction) {
>>       |                                 ^
>>
>> Signed-off-by: Yonggang Luo <luoyonggang@gmail.com>
>> ---
>>  tests/test-replication.c | 4 ++++
>>  1 file changed, 4 insertions(+)
>>
>> diff --git a/tests/test-replication.c b/tests/test-replication.c
>> index e0b03dafc2..9ab3666a90 100644
>> --- a/tests/test-replication.c
>> +++ b/tests/test-replication.c
>> @@ -554,6 +554,9 @@ static void sigabrt_handler(int signo)
>> 
>>  static void setup_sigabrt_handler(void)
>>  {
>> +#ifdef _WIN32
>> +    signal(SIGABRT, sigabrt_handler);
>> +#else
>>      struct sigaction sigact;
>> 
>>      sigact = (struct sigaction) {
>> @@ -562,6 +565,7 @@ static void setup_sigabrt_handler(void)
>>      };
>>      sigemptyset(&sigact.sa_mask);
>>      sigaction(SIGABRT, &sigact, NULL);
>> +#endif
>>  }
>> 
>>  int main(int argc, char **argv)
>>
>
> This is already fixed by a patch from Thomas.

Well, my patch was to simply disable test-replication on Windows ... if
it is working with this modification here, that's certainly better than
disabling it.

 Thomas