Re: Test release for GSL 2.8

[Brijesh Upadhaya]

Hi,

All tests passed for gsl-2.7.1+ in Debian 12 with GCC-12.2.0 and glibc-2.36.

Maybe you should update the author names in the source codes inside gsl/linalg. 
It seems that you have added/updated a few routines. 

I don't know if this is necessary, but the scaling of a complex (sparse) matrix 
with a real vector, complex sparse matrix-vector multiplication, and 
unpacking(in place) a complex vector could be added. I had to somehow do the 
above-mentioned stuff while solving a time-harmonic FE problem. 

There are alloc and calloc for the blocks and vectors. Is there any particular 
reason for not having anything to resize a block or a vector?

In gsl/spmatrix/getset_source.c, _set() replaces the matrix elements, could it 
be possible to assemble the matrix as

*(BASE*) ptr = *(BASE*) + x; 

and copy the entire code to a new name.


Regards, 
Brijesh

> On 11. May 2024, at 1.18, Mohammad Akhlaghi <makhlaghi@gmail.com> wrote:
> 
> Dear Patrick,
> 
> Thank you very much for sharing the good news about the new GSL release.
> 
> I just configured, built and checked the tarball on an Arch GNU/Linux with 
> GCC 14.1.1 and Glibc 2.39 (on 12 threads to speed it up):
> 
> ./configure && make -j12 && make check -j12
> 
> There were no warnings or errors in the compilation or in the checks! 
> Everything went smoothly all the way until the end :-).
> 
> Thanks for all the work on this wonderful package; GSL is a mandatory 
> dependency of GNU Astronomy Utilities; and we rely heavily on it. I look 
> forward to the release of GSL 2.8.
> 
> Cheers,
> Mohammad
>