Re: [Hamlib-developer] What am I missing?

[Greg T. <gd...@le...>]

Nate Bargmann <n0...@n0...> writes:

> * On 2025 26 Jul 09:52 -0500, George Baltz wrote:
>> I'm wondering if this is why shortfreq_t is still a long.  Signed 16 bits
>> might be too short (repeater offset) but 32 bits would work.
>> 
>> Would it be possible to change shortfreq_t to int32 for 5.0?
>
> I think so as we're communicating with the advancement of the major
> number that API changes are made and ABI is broken.
>
> Any such values that appear to be limited should probably be evaluated
> for 5.0.

And implicitly, if there's any concern about the actual size of the
type, C99(?) fixed-width types like int32_t are prepared.

I am surprised though about 16-bit ints.  Perhaps that's a Windows thing
semi-recently.  Unix on the PDP-11 (V6, V7 and 2.8-2.11 BSD) had 16 bit
ints and 32 bit longs.   I am pretty sure the Vax had 32 bit int right
away.  And that's mid 80s with PDP-11s fading by mid 90s.