Re: [PATCH] Skip tests requiring "alignof (void)" when compiling using clang

[Pedro Alves <pedro@palves.net>]

On 7/2/20 9:52 PM, Simon Marchi wrote:
> On 2020-07-02 4:49 p.m., Pedro Alves wrote:
>> I think this should still test GDB's support.  And a
>> comment would be helpful.  Like:
>>
>> # As an extension, GCC allows void pointer arithmetic, with
>> # sizeof(void) and alignof(void) both 1.  GDB supports GCC's
>> # extension.  Clang does not.
>> if ![test_compiler_info clang*] {
>>     set expected [get_integer_valueof a_void 0]
>>     gdb_test "print alignof(void)" " = $expected"
>> } else {
>>     gdb_test "print alignof(void)" " = 1"
>> }
> 
> Indeed.  Otherwise, let's say that clang gain this feature, it would remain untested
> (when testing with clang) and nobody would think of coming here to update the test
> case.  Here, if clang gains the feature, then I suppose it would generate a FAIL,
> which would prompt us to update the test case.

Actually, if Clang gains the feature, the test would still pass.
GDB's alignof support is builtin, does not depend on what the compiler
outputs.