Re: CVS commit: src/sys/arch/mac68k/mac68k

[christos%astron.com@localhost (Christos Zoulas)]

In article <21181.1550538876%splode.eterna.com.au@localhost>,
matthew green  <mrg%eterna.com.au@localhost> wrote:
>Christos Zoulas writes:
>> In article <20190219003451.093B2FB17%cvs.NetBSD.org@localhost>,
>> matthew green <source-changes-d%NetBSD.org@localhost> wrote:
>> >-=-=-=-=-=-
>> 
>> >+	memcpy(intrnames, inames, MAX_INAME_LENGTH);
>> 
>> That should be:
>> 	memcpy(intrnames, inames, sizeof(intrnames));
>
>well.
>
>maybe.  but it's not really any better.  intrnames is not
>defined with MAX_INAME_LENGTH so using it in any place is
>still a hack...
>
>ie, the length is actually eintrnames-intrnames, but i am
>not aware of a way to do that in pure C (leaving the names
>as they are now, which they're needed for vmstat.  it needs
>'eintrnames' to know the end of the list of strings.)
>
>i'd like to replace it with a much less gross setup but i
>didn't see one...
>
>and the good thing is, gcc will whine if it grows ;)

Yes, the good thing about the sizeof(intrnames) is that if the type
changes it is still right...

christos