Hi Przemek,
> On Tue, 09 Feb 2010, [email protected] wrote:
>
> Hi Viktor,
>
>> 2010-02-09 15:43 UTC+0100 Viktor Szakats (harbour.01 syenar.hu)
>> + Added '( HB_SIZE )' casts to 'strlen()' calls where
>> it was converted to HB_SIZE. This pacifies long standing
>> win64 warnings.
>> Remaining warnings are (among others) useful to spot places where
>> non-HB_SIZE is used in Harbour code to store sizes.
>
> This modification was not necessary.
> When we complete type modification then we switch HB_SIZE in WIN64 builds
> to 64bit integer. It will pacify all such warnings automatically.
> And of course will nicely show all places where we have not casted
> HB_SIZE -> [U]LONG conversion which we can easy locate and fix.
No problem, all these locations are easily found
using 'grep "( HB_SIZE ) strlen("' along with all
those which were already cast. I had reversion in
mind, and the above patch was a complete one for
strlen(), I don't plan to add casts for more
obscure places. (ops hbmemio/hbnetio has 1-1
exceptions... I'll revert those in next commit)
Until then, it's easier to focus on warnings showing
real problems (some code still use size_t, int and
HB_ULONG, even HB_USHORT to store sizes).
BTW, we've been discussing it, but haven't finally
decided about HB_SIZE size. Should it be tied to
platform integer size, or we roll our own. Anyway
I agree to use what the platform offers.
So, what to finally aim for: HB_SIZE <- HB_MAXINT?
Brgds,
Viktor
_______________________________________________
Harbour mailing list (attachment size limit: 40KB)
[email protected]
http://lists.harbour-project.org/mailman/listinfo/harbour
