This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [committed] Fix MIPS -mint64 deprecation check


Eric Christopher <echristo@redhat.com> writes:
>> 
>> 	* config/mips/mips.c (override_options): Only warn about -mint64
>
>>  
>>    /* Deprecate -mint64. Remove after 4.0 branches.  */
>> -  if ((target_flags_explicit & MASK_INT64) != 0)
>> +  if (TARGET_INT64)
>>      warning ("-mint64 is a deprecated option");
>>  
>
> Not sure I agree actually. If you use it on the command line you should
> get a warning that it's deprecated because you actually typed it on the
> command line there and deserve the warning that what you typed won't be
> supported.

But as per the covering note, target_flags_explicit doesn't (and isn't
supposed to) tell you whether -mint64 was used on the command line.
It just means that _some_ option explicitly set the value of MASK_INT64.
In the case of -mlong32, that's not the same thing at all.

I tried to justify why things like "-mint64 -mlong32" (where MASK_INT64
is first set and then cleared) aren't worth worrying about.  If you
disagree strongly, we could add a spec string like %{mint64:%e....}.
I think that's the only way of knowing for sure whether -mint64 was
used and later revoked.  (Note that there's no -mno-int64; -mlong32
is the only way of cancelling an earlier -mint64.)

Richard


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]