[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
dim_vector constructor problem
From: |
John W. Eaton |
Subject: |
dim_vector constructor problem |
Date: |
Wed, 28 Apr 2010 01:53:27 -0400 |
The changeset
http://hg.savannah.gnu.org/hgweb/octave/rev/69ecfbffcf4f
is causing trouble with builds that use 64-bit indexing
(--enable-64). There is a bug report about it:
https://savannah.gnu.org/bugs/?29692
The real cause is that it introduces an ambiguity for calls like
dim_vector (0, 0)
when octave_idx_type is not a typedef for int.
I thought I could solve the problem with a few strategically placed
casts, or by making the second argument of the new constructor be int
instead of size_t, but that did not work.
Given that we also now have
dim_vector (i, j, k, ...)
constructors for up to 7 dimensions, does anyone object to removing
the
dim_vector (const octave_idx_type *vec, size_t vec_size)
constructor that was added by this changeset? To me, it seems the
simplest solution to the problem. Do we lose anything by not having
this constructor now?
I checked in a changeset to remove the constructor and I closed the
bug report. But if there are objections, I can reconsider, provided
that there is also some reasonable way to solve the 64-bit compile
problem.
jwe
- dim_vector constructor problem,
John W. Eaton <=