|
Boost Users : |
From: Joaquín Mª López Muñoz (joaquin_at_[hidden])
Date: 2006-04-18 03:20:53
Patel Priyank-PPATEL1 ha escrito:
> Hi Joaquin,
>
> First of all thanks for your time and response. I have it working as follows, but seems I have to recalculate iterators if in case I later on modify group associated with that access point. Is there any way that iterators will know automatically update when I update access point in future to some other group?
> Let me know if you have come across that one. I am going to look in your code in detail and get back to you in case of any questions.
Hello Priyank, there's an issue in your code:
> // change paging group from access point 4
> a4->pg(2);
This is incorrect and can lead to hard-to-detect failures later on during the execution
of the program: basically, you're changing an element key "under the table", that is,
without the multi-index container having a chance to notice, which directly
breaks the internal invariants the container relies on. Under usual conditions it
is difficult to change a key in such a way, because the elements of a multi-index
container are treated as const, but here you can easily get by because it is
pointers that you are storing.
The proper way to modify an element is by using replace, modify of modify_key,
as explained in
http://boost.org/libs/multi_index/doc/tutorial.html#ord_updating
In your case, instead of what you've got you could write the following:
// get an iterator to a4. You could also have stored those iterators in the
// first place when inserting the elements.
Access_Points::iterator it=pool.find(4);
// change paging group from access point 4
pool.modify(it,boost::bind(&Access_point::pg,_1,2));
By using modify(), the multi-index container is on control of the updating
operation, and will rearrange a4 according to its new paging group.
> // NEED TO RECALCULATE ITERATORS
> boost::tuples::tie(pg0, pg1) = get<paging_group> (pool).equal_range(1);
As for your original question, yes, you've got to recalculate these iterators in
the general case, as elements might have been moved around, inserted or
deleted: for instance, if the element originally pointed to by pg0 had been
deleted, pg0 would not even be valid after that.
Joaquín M López Muñoz
Telefónica, Investigación y Desarrollo
Boost-users list run by williamkempf at hotmail.com, kalb at libertysoft.com, bjorn.karlsson at readsoft.com, gregod at cs.rpi.edu, wekempf at cox.net