On Thu, 19 Apr 2012, Kirk, Benjamin (JSC-EG311) wrote:
1812 roystgnr * This function is now deprecated - write the
1812 roystgnr * libmesh-devel mailing list if you need it reimplemented.
598 benkirk */
598 benkirk void delete_system (const std::string& name);
So here we are – 4,000 svn revisions later, the method is still in there and
I’m dubious that it will work right if it is called.
Roy, do you happen to remember what led up to the deprecation?
I was sorely tempted to make this question the lead-in to a hilarious
paragraphs-long joke, but a simple "no" is probably kinder.
I suppose asking the list to reimplement it would be an exercise in
futility, but if we can figure out what the issues were I might be
able to do something about it...
My best guess:
This came right after r1811, "Made EquationSystems::solve() operate in
numeric order", intended to standardize that as a way for doing a
sweep through a decoupled problem. But the trouble with accessing
systems as if they're a vector is deciding semantics for when a
deletion makes the vector sparse.
Do the remaining systems (and every related DoFObject, etc) get
renumbered? Seems too hard to keep that straight, but otherwise we
need to go through every bit of code that's using indexed systems and
make sure it can handle the NULL case. We also need a pointer-based
version of get_system<T_sys>(uint), or at least a bool-returning
has_system(uint), so that code can actually *detect* a NULL case.
Does n_systems() still return the number of systems? That seems to be
the most straightforward interpretation, but then we need an
end_sysnum() to return the largest index for when that's what's really
relevant, and we need to make sure any loop over indices is using that.
Those are the first issues that come to mind.
---
Roy
------------------------------------------------------------------------------
For Developers, A Lot Can Happen In A Second.
Boundary is the first to Know...and Tell You.
Monitor Your Applications in Ultra-Fine Resolution. Try it FREE!
http://p.sf.net/sfu/Boundary-d2dvs2
_______________________________________________
Libmesh-devel mailing list
Libmesh-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/libmesh-devel
