[389-devel] Re: Use of int types in the code base,

Hi William!
Thank you for the email. It has clarified the things for me. :)
I still have one question though.

Do I understand right that I need also to cast types with these types from inttypes.h?
So use it not only in the defenitions.
- sprintf(buf, "%lu", (long unsigned int)maxsize);
+ sprintf(buf, "%" PRIu64, (uint64_t)maxsize);

Thanks,
Simon

----- Original Message -----
> From: "William Brown" <william@blackhats.net.au>
> To: "389-devel" <389-devel@lists.fedoraproject.org>
> Sent: Thursday, March 8, 2018 12:42:03 AM
> Subject: Use of int types in the code base,
>
> Hi there,
>
> http://www.port389.org/docs/389ds/development/coding-style.html#types
>
> In a few reviews I still see this sometimes.
>
> It's pretty important that we keep moving our quality standard higher,
> and having known type sizes is important to this. Types like int and
> long are unknown sizes until you compile it depending on platform.
>
> As a result, it's really important we use the intX_t and uintX_t types
> so we have guarantees of our values. I would encourage the use of
> int64_t and uint64_t, because while they are "larger", it's
> significantly faster for a modern 64bit system to process these values
> than their 32bit counterparts.
>
> Another note is that arrays index by size_t, not 'int', so we should
> always keep this in mind.
>
> Finally, because we are using c99 now, this means we should avoid:
>
> size_t i = 0;
>
> for (i = 0; i < cond; i++) {
> ...
> }
>
> When we really should scope our values. Scoping is good because it
> limits possibility of data corruption to flow and other mistakes such
> as re-use of values. This means:
>
> for (size_t i = 0; i < cond; i++) {
> ...
> }
>
> Thanks!
>
> --
> Thanks,
>
> William Brown
>
_______________________________________________
389-devel mailing list -- 389-devel@lists.fedoraproject.org
To unsubscribe send an email to 389-devel-leave@lists.fedoraproject.org