On 11-08-21 06:21 PM, Daniel Kahn Gillmor wrote:
> Hi good slony folks--

Thanks for the report.

>
>
>   0) overflow isn't being handled properly.  The simplest way to fix this
>   is probably to do this kind of internal computation in floats, and then
>   cast back to integers later.

Or long long ?


>
> some secondary issues worth considering are:
>
>   1) this is a 64-bit architecture -- why aren't these using 64-bit
>   integers instead of 32-bit?  Maybe there are some odd compiler flags in
>   use?
>
>   2) should these ints be unsigned instead of signed?
>
> I hope this problem report is useful!  I'd be happy to answer any
> questions or to clarify concerns other people have.  Would it be useful
> to report this concern anywhere other than the mailing list?

Reports on the mailing list tends to get attention. I have put this 
report into a bug-report at 
http://www.slony.info/bugzilla/show_bug.cgi?id=235 to track it.



>
>     --dkg
>
>
>
> _______________________________________________
> Slony1-general mailing list
> Slony1-general at lists.slony.info
> http://lists.slony.info/mailman/listinfo/slony1-general