This week I almost finished BLAS, it only needs some more documentation and fixing a few tests. I also added new wrappers to Matrix which let the user set values of a matrix row by row (set_row) or column by column(set_col), output a column matrix(col), output a row matrix(col) and I made modifcations to the wrapper as_list that Jonathan already wrote which output the matrix as a list. I added more documentation to Histogram. Finally I took a look at other modules I could work on this week with our current tyepmaps. I chose the following modules to work on this week : Multifit, Interp, Wavelet and NTuple. I'll also take a look at Fit but the only function I tried to test died so I might have to play with the current typemaps.
This week I added examples to the documentation of all subsystems usable so far. This was a missing part for a release to CPAN. I talked with Jonathan about a date we should release and the 20th of each month would be a good date since I'll have some time to polish the release for this month and next month release would a be released near the end of gsoc.
This week I also added some Object-oriented wrapper to the Matrix subsystem. I'm starting to understand better the object-oriented concept which is something I haven't yet seen in my university courses. Finally, I documentation and test functions which don't need the gsl_vector_float structure to the BLAS subsystems. Level-1 function are now almost all functional and Level-2 functions are in progress.
Next week I'll work on finishing BLAS and Histogram testing and documentation, adding more tests to CBLAS and I'll review some function we couldn't make work before since Jonathan seems to have solved the issue with the typemap for inputing arrays. The Sort subsystem is harder than I tought and I'll also have a look back on it.
This week I worked on adding more documentation to some subsystems and fixing some failing tests. With the help of Jonathan I've been able to fix or mark as TODO all the failing tests so we could merge the bleed repository into the master one. I also added a lot of documentation, mainly to Linalg, Randist, CDF and Permutation. There is still need for more documentation on many subsystems and I plan to work on that in the following weeks.
Jonathan told me that the problems we have with typemaps make it hard to follow the timeline. Instead I'll work on subsystems that I feel can be made with our current understanding of timemaps. For the next week it means working on CBLAS, NTuples and Sort subsystems. CBLAS already has a test suite in GSL that I'll probably use. The BLAS subsystem however seem to have a problem with the type 1 functions. It ask for gsl_vector_float datatype as argument but the gsl_vector_float functions seems to be deprecated. I plan to write to the GSL developpers to know what's going on.
We should put our first release on CPAN soon, we only need to polish a few more things.