Re: clang-10 warning in hash.c

[Tim Rühsen]

Hi Bruno,

On 1/27/20 2:09 PM, Bruno Haible wrote:
>> Not sure if the compiler is correct here, but maybe worth a look:
>>
>> hash.c:549:11: error: implicit conversion from 'unsigned long' to
>> 'float' changes value from 18446744073709551615 to 18446744073709551616
>> [-Werror,-Wimplicit-int-float-conversion]
>>       if (SIZE_MAX <= new_candidate)
>>           ^~~~~~~~ ~~
>> /usr/include/stdint.h:227:22: note: expanded from macro 'SIZE_MAX'
>> #  define SIZE_MAX              (18446744073709551615UL)
>>                                  ^~~~~~~~~~~~~~~~~~~~~~
> 
> This warning is pointless, because
>   - Since the next float below 18446744073709551616 = 0x10000000000000000
>     would be                   18446742974197923840 =  0xFFFFFF0000000000
>     the comparison result is the same for the two values ...615 and ...616.
>   - The compiler inserts the implicit conversion only because of the '<='
>     operator.
> IMO you should file a ticket with the clang people.
> 
> Inserting a cast to 'double'
> 
>   if ((double) SIZE_MAX <= (double) new_candidate)
> 
> would not help, because
> the next double-float below 18446744073709551616 = 0x10000000000000000
> would be                    18446744073709549568 =  0xFFFFFFFFFFFFF800

Thanks for the understandable explanation !

Sadly, bugs.llvm.org disabled self-registration. So they won't get a bug
report from me (looks like they want to encapsulate from the rest of the
world). Instead -Wno-implicit-int-float-conversion will be added to the
clang options.

Regards, Tim

signature.asc
Description: OpenPGP digital signature