[Rpm-maint] recognizing epoch, take two

Ales Kozumplik akozumpl at redhat.com
Wed Nov 16 11:32:52 UTC 2011

On 11/16/2011 11:20 AM, Michael Schroeder wrote:
> On Wed, Nov 16, 2011 at 11:17:23AM +0100, Ales Kozumplik wrote:
>> I use long int type to pass the epoch around. This type can both cover the
>> entire range of RPM_INT32_TYPE (the type of epoch value) and -1 for the
>> 'any epoch' semantics. Compared to the previous version, it saves us one
>> malloc and is not longer.
> Hmm, is sizeof(long int)>  4 true for all platforms?
> Cheers,
>    Michael.


You are right, thank you for catching this. I will use int64_t in the 
final patch.


More information about the Rpm-maint mailing list