(from the boost-user group) ... is ... I don't know why any of the ublas devs have commented on this at boost-user (or maybe they have but I missed it?). There...
It looks like support of exact types in uBLAS is not complete: ============ namespace ublas = boost::numeric::ublas; int main() { typedef ublas::vector<int>...
I have created a new branch in CVS called 'uBLAS_pure'. Effectively this is a large clean up of the current uBLAS without any of the old compiler workaround....
Hello all, I used uBLAS_pure on IBM. After the modifications I did yesterday on my local version to get old stuff working, everything seems to work well on ...
Dear All, Looking at the current sparse matrix documentation it seems that the type descriptions for sparse, compressed and coordinate are highly cryptic. I am...
There was some positive response on my proposal to organise a ublas developers meeting. So lets move on and let us try to determine where and when to organise...
Hi Karl, Fres, Peter, Paul, In terms of naming how about: size_type nnz_filled (); size_type nnz_capacity (); This uses existing naming (nnz) with regard to...
Gunther Winkler wrote ... cd boost/numeric/ublas cvs update -r uBLAS_pure Note that using the '-r' flag to update implies a 'Sticky Tag'. Thus it remembers...
Hi Gunter, I changed over to std::stable_sort and ran into some problems with the arrays being corrupted. In one test case I have the follow input data I1 I2 V...
Dear all, I use in my program fixed size arrays, what is not yet supported by uBlas directly. Anyway, I do not mind if the size-tests are on compile-time or on...
Dear All, I stumbled across and fixed a couple of nasty bugs. First bug: There was an inconsistency in the the non_zero() function return values. The...
While thinking about the semantics for the insert operation I have been trying to come up with a better name. At the moment we have: set_element - my...
For some practical arrangements, I plan on contacting collectively all persons that have added their name on the dev-meeting page on the wiki to avoid making...
Hi Toon, Robbie, ... I think I have been once again rather confusing in my replies. I think long term we will need to define the Sparse concepts. The most...
I fundamental question I would like to ask is: Do objects of type value_type in the ublas containers need to be zero-initialisable. In many discussions people...
Dear All, I have committed a major set of changes to the UBLAS_pure branch. This implements some major changes to element assignment, access and proxies. This...
Currently the sparse matrices are a model of the matrix-concept ? Should'nt we develop a sparse-matrix concept that is a refinement of the matrix-concept. For...
I would like to propose that we remove the member function 'insert', 'erase' and 'clear' from all models of vector-concept (such as ublas::vector). The reason...
Resize: ======= The vector-concept documentation and vector model documentation mentiona resize(new_size,t) function. However in the real world there is ...
(to all except Miceal: we are currently trying to improve the documentation of ublas. This work is currently done in the ublas cvs (the last reason of...
Micheal, is is possible that you forgot to checkin the container_concept page because the index.html points to it but the link is dead (I'm planning on working...
Currently following persons have confirmed: # Knapen Toon # Meerbergen Karl # Stevens Micheal # Winkler Gunther # McCulloch Ian If you're planning on...
Just committed a couple of changes. I have renamed the concept documents '*_concept.htm' to make them easier to identify. I have merged range and slice. Toon I...
Dear all, cureently I'm in the process to include support for an ibm SP3 again, AIX 5.2, xlC v6. I have the follwoping problem, that I get lots of error...
Dear All, Both in reading and write the uBLAS documentation the class names sparse_matrix and sparse_vector make things very confusing. In the past I have...
Matthias, I have committed a fix. The problem was in unbounded_array. In order to default construct the allocated memory I was using an array placement new....
