Re: [PATCH v2] Class-ify ptid_t

[Simon Marchi <simon.marchi@ericsson.com>]

On 17-04-07 06:48 AM, Pedro Alves wrote:
> On 04/07/2017 10:25 AM, Philipp Rudo wrote:
> 
>> In this context I ask myself what the error message is supposed to
>> say (unfortunately you are the first one in GDB actually using it, all
>> others only use an empty string).  Is it what's expected ...
>>
> 
> That's because the static assertion failure text includes both a
> file:line indicating where's the failure, and modern gcc/clang
> show you the line that failed, so it's not that big of a deal to
> not include a message:
> 
>  src/gdb/thread.c:1589:1: error: static assertion failed: 
>   static_assert (1 == 0, "");
>   ^
>  $
> 
> (and in C++17, static_assert got an overload with no second
> parameter, even.)
> 
>>
>> static_assert (pid == ptid_t (1, 0, 0), "pid operator== returned false")
>>
>>
>> I tend more to what is expected.
> 
> Definitely it should say what is expected, just like regular
> testsuite tests.  Or say nothing when the expression is pretty clear
> already, which works for me too.  :-)

Heh, why did you not say that earlier, it would have saved me some time ;)

@Philipp: To come back to the verb tense thing.  Now that I think of it, when
I read the lines, I mentally insert a "that" in there.  So

  static_assert (!pid.matches (lwp), "pid doesn't match lwp");

becomes mentally

  assert that pid doesn't match lwp

I think it comes from using jUnit's assertThat:

https://github.com/junit-team/junit4/wiki/matchers-and-assertthat

Simon