CGAL users discussion list

[郑银河 <>]

Hi Marc,

Thank you very much for your reply. It seems that I have found a way to do the job based on your idea. Let me try to implement it.

Thanks again

Yinhe

2016-03-03 1:43 GMT-07:00 Marc Glisse <>:

On Wed, 2 Mar 2016, 郑银河 wrote:

I used the exact kernel in my project to do the exact computation. However,
it takes a lot of time to calculate some intermediate results. I am finding
a way to restore these results, which are represented using the exact
number type CGAL::Exact_predicates_exact_constructions_kernel::FT, to the
hard disk "as is“ in an exact way and read these data from the disk before
carrying out another calculation. Is there any way to do this kind of job
in CGAL? If not, (which is very likely to be the answer of the previous
question because I tried very hard to search the manual) how can I restore
these intermediate results for another calculation?

Epeck is lazy, it stores a tree of operations that was used to produce this object and an approximation. That seems hard to serialize directly. You could use x.exact() on each number
http://doc.cgal.org/4.7/Number_types/classCGAL_1_1Lazy__exact__nt.html#a1b1e71678b9cb5dcfbca923e6abb0959
this will return a rational number that can be serialized. But if you do it all the time, the laziness becomes useless, and you might as well define CGAL_DONT_USE_LAZY_KERNEL, so FT will directly be that rational type.

--
Marc Glisse

--
You are currently subscribed to cgal-discuss.
To unsubscribe or access the archives, go to
https://sympa.inria.fr/sympa/info/cgal-discuss