Re: clang-10 warning in hash.c

[Paul Eggert]

On 1/27/20 5:09 AM, Bruno Haible wrote:
> IMO you should file a ticket with the clang people.

That sounds appropriate.

In the past I've worked around problems like these by writing things 
like "SIZE_MAX + 1.0f <= new_candidate" instead of "SIZE_MAX <= 
new_candidate" (partly under the argument that this better matches the 
mathematical comparison we're actually trying to do), but it sounds like 
Clang 10 would warn about that too.

> Inserting a cast to 'double'
>
>   if ((double) SIZE_MAX <= (double) new_candidate)
>
> would not help

Presumably the Clang folks want us to insert a cast to 'float', e.g., 
"(float) SIZE_MAX <= new_candidate". However, I dislike casts because 
they're too powerful.

If there's no way to pacify Clang without casting, then I suggest 
disabling the warning instead. In GCC, we already disable Clang's 
-Wswitch, -Wpointer-sign, -Wstring-plus-int and -Wunknown-attributes 
warnings, and perhaps after Clang 10 comes out we'll disable 
-Wimplicit-int-float-conversion too. Some diagnostics are just too much 
trouble to pacify.