octave-maintainers
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Array<T> cleanup?


From: John W. Eaton
Subject: Array<T> cleanup?
Date: Mon, 13 Oct 2008 11:33:51 -0400

On 13-Oct-2008, Jaroslav Hajek wrote:

| 1. what is the purpose of Array<T>::set_index and stuff? It seems to
| me that it is only used in the assign functions. Perhaps a relic of an
| old attempt for lazy indexing?

I think the motivation for doing indexed assignment this way was that
way back in the dark ages, the location where the index was available
was different from the place where it was needed, so I decided to push
the indices on to the Array object.  I never really liked this
solution, but it worked and solved whatever problem it was that I was
facing at the time.

| Other indexing functions (index and maybe_delete_elements) do not use
| Array<T>::idx but pass the indices as arguments. Unless I'm
| overlooking something, I'd suggest making the assign functions member
| functions and pass the indices as arguments.

Yes, this is something that I've been wanting to do for some time, so
if you would like to do it, please go ahead.  But for this change,
will you please post the changeset first before committing?

| 2. There seems to be several of orphaned header files in liboctave,
| namely Bounds.h, Bounds.cc, LinConst.h, LP.h, QP.h, NLP.h FEGrid.h...
| these header sources are included in build process (sometimes only
| headers), but not used anywhere.
| qp and sqp are provided by m-files, linear programming by glpk. And I
| intend to supply an m-file implementation of fsolve, that will make
| MINPACK obsolete, along with NLEqn.h, NLEqn.cc and NLConst.h. FEGrid
| seems unrelated to anything in Octave.
| Is there any reason to keep these files around?

If they are really not used, then I dont' have any objection to
removing them.

Thanks,

jwe


reply via email to

[Prev in Thread] Current Thread [Next in Thread]